Skip to main content

Full text of "Occasional papers of the California Academy of Sciences"

See other formats




OCCASIONAL PAPERS 

OF THE 

CALIFORNIA ACADEMY OF SCIENCES 

No. 149 September 2 1,2000 

EMBIA 

Contributions to the Biosystematics 

of the 
Insect Order Embiidina 

Part 1 
Origin, Relationships and Integumental 
Anatomy of the Insect Order Embiidina 

Part 2 
A Review of the Biology of Embiidina 

By 
Edward S.Ross 



Published by the California Academy of Sciences 
San Francisco, California 



..J.^ - \.' •■"'■ 







EMBIA 

Contributions to the Biosystematics of the 
Insect Order Embiidina 




Male Cloihoda n. sp., (in alcohol, body length 20 mm), related to C. longicauda Ross, of Peru, second only 
to C. nobilis (GcTsi) as the most plesiomorphic species of the order. Plesiomorphic features include: large 
size, complex wing venation and almost completely symmetrical terminalia, including exceptionally long 
cerci. Specimen from upper Napo River region of Ecuador (see Figs. 39A and 40A). 



OCCASIONAL PAPERS 

OF THE 

CALIFORNIA ACADEMY OF SCIENCES 



No. 149 



EMBIA 

Contributions to the Biosystematics 

of the 
Insect Order Embiidina 

Part 1 
Origin, Relationships and Integumental 
Anatomy of the Insect Order Embiidina 

Part 2 

A Review of the Biology of Embiidina 

By 

Edward S. Ross 




Published by the California Academy of Sciences 
San Francisco, California 



SCIENTIFIC PUBLICATIONS 

Alan E. Leviton, Editor 
Katie Martin, Managing editor 

© 2000 California Academy of Sciences, Golden Gate Park, San Francisco, California 941 18 

All rights reserved. No part of this publication may be reproduced or transmitted in any form or 
by any means, electronic or mechanical, including photocopying, recording, or any information 
storage or retrieval system, without permission in writing from the publisher. 

ISSN 0068-5461 



Table of Contents 
Part 1. Origin, Relationships and Integumenta! Anatomy of the Insect Order Embiidina 

Preface 1 

Acknowledgments 1 

Origin and Relationships 

Summary 3 

Relationships 3 

Origin 3 

Fossil record 4 

Pretertiary fossils 5 

Tertiary fossils 6 

Integumental Anatomy 

Introduction 7 

Methods 7 

Head 8 

Eyes II 

Antennae II 

Mouthparts 12 

Cervix 13 

Prothorax 15 

Pterothorax of males 16 

Meso- and metathorax of females 18 

Legs 21 

Prothoracic legs 21 

Nature of embiid silk 23 

Mesothoracic legs 23 

Metathoracic legs 24 

Wings 25 

Venation 27 

Wing pigmentation 30 

Wing expansion following ecdysis 33 

Wing articulation 33 

Flexion of wings 34 

Flight 34 

Reduction and elimination of wings 34 

Abdomen 38 

External genitalia of females 39 

Internal genitalia of males 39 

External genitalia of males ("Terminalia") 40 

Anomalous male terminalia 46 

Portfolio of male terminalia in Embiidina 47 

Literature Cited 52 

Part 2. A Review of the Biology of Embiidina 

Summary I 

Methods 1 

General biology 1 

Social behavior 8 

Behavior of adult males 9 



Color of adult males 10 

Mating 11 

Eggs and their protection 13 

Development 15 

Expansion of wings 16 

Parthenogenesis 17 

Diet 18 

Movement 20 

Habitats 

Tropical Evergreen Forests 21 

Tropical Cloud Forests 22 

Seasonally-Dry Grassy Woodland 22 

Semi-Arid, Open Grasslands 24 

Desert Areas 25 

Human Habitats 25 

Geographic range 28 

Ecological range 28 

Natural dispersal 29 

Dispersal by man 29 

Natural hazards 

Predators 31 

Ectoparasitoids 32 

Endoparasitoids 34 

Egg Parasitoids 34 

Pathological Hazards 34 

Literature Cited 35 



Part 1 

Origin, Relationsliips and Integumental 
Anatomy of the Insect Order Embiidina 



Preface 

This is the first issue of EMBIA, a new pubhca- 
tion which, at least initially, will serve as the outlet 
for my large number of long-shelved manuscripts on 
Embiidina. Many of these accumulated because I 
had intended them to be chapters in a series of mono- 
graphic volumes. However, at this late stage in my 
life, this will not be possible. 

Over the years, of necessity, laboratory research 
was interrupted by expeditions — one lasting eighteen 
months — to assemble a meaningful collection of the 
order. Besides my curatorial duties at the California 
Academy of Sciences. I was. of course, distracted by 
many other activities, such as making available to 
popular and educational media, the rich collection of 
nature photographs I made during expeditions. In- 
come from such activity paid much of the cost of the 
fieldwork and, now, a portion required to produce 
EMBIA. 

Recent availability of desktop methods of pub- 
lishing has encouraged me to discontinue scattering 
my contributions in a diversity of periodicals. If I 
live long enough (2000 is my 85th year!), many is- 
sues of EMBIA — mostly on systematics — will ap- 
pear in rapid succession. 

I am using Embiidina (Embidina, Hagen, 1861), 
the first group name applied exclusively to embiids, 
as the order's name in preference to more recent 
Embiodea (Kusnezov, 1903) and Embioptera 
(Shipley. 1904). Embiidina has had important early 
use in monographs and is now rapidly supplanting 
the awkward, inappropriate name Embioptera. 

Acknowledgments 

These will briefly serve this and future issues of 
EMBIA. This anatomical treatment is in part based on 
my unpublished Ph.D. dissertation (U.C. Berkeley, 
1941). During this early work, I benefited from many 
contacts with the late G. P. Ferris of Stanford Univer- 
sity, as is evident in the style of my drawings. Follow- 
ing four years of Army service in World Wai' II, during 



which I collected embiids in New Guinea and the Phil- 
ippines, I began a series of extensive embiid surveys in 
all major centers of the order's evolution. The result- 
ant collection, now numbering more than 300,000 
specimens, representing about 1000 species (75% 
undescribed) is almost overwhelming. 

In this essential fieldwork I was helped by hun- 
dreds of individuals, too numerous to mention. Ma- 
jor financial support, in addition to my own funds, 
came from the John Simon Guggenheim Founda- 
tion (Amazon Basin), the National Geographic So- 
ciety (Africa and Indo-Australia), the National 
Science Foundation (field and laboratory activities), 
the Dietz Fund, San Francisco Foundation (Africa), 
the Schlinger Foundation and Health and Habitat. 

Special thanks should be given to my field assis- 
tants, who, like myself, participated without salary. 
They were: Wilda S. Ross, Sandra M. Ross, Eveit I. 
Schlinger (second South American trip), Robin 
Leech (first African trip), David Cavagnaro and 
Clark Ross (Indo-Australia), Kenneth Lorenzen (sec- 
ond African trip). Alan Stephen (third African trip) 
and Mike Irwin (fourth African trip). 

Later Robin Leech collected cultures for me es- 
pecially in southeastern Asia and David Cavagnaro 
did the same in Central America and Ecuador. 

At the California Academy of Sciences, hundreds 
of embiid cultures were maintained by the late Peter 
Rubtzoff. Currently, with a Schlinger Foundation 
grant and my personal funds, the Entomology 
Department's very able secretary, Julieta Parinas, as 
off-duty employment, is typing manuscripts for 
EMBIA and preparing pages for publication. Ear- 
lier, also with Schlinger funds. Andrei Sourakov, as- 
sisted me in many ways. I am also greatly indebted 
to Jarmila Kukalova-Peck of Carleton University, 
Canada for valuable information on wing, 
articulation. venation and Paleozoic relationships of 
embiids. Norman D. Penny, and Katie Martin, Cali- 
fornia Academy of Sciences, kindly checked manu- 
scripts and made editorial improvements. 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



# 



^ 




f' 



..# 



Figure A. Except for size and coloration, all female embiids have similar appearance. This photograph clearly 
shows the spinning foretarsi and the enlarged hind femora within which large tibial flexor muscles motivate quick 
reverse movement. New genus and species, Oligotomidae. Thailand. Body length 12.0 mm. 




4-' ^r - -V 

Figure B. Typical adult male. Like females, alate males are similar in appearance throughout the order. 
Antipaluria caribbeana Ross. Clothodidae. Venezuela. Body length 13.0 mm. Frontispiece shows spread 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



Origin and Relationships 



Summary 

No proven pre-Tertiary embiid fossils have been 
found but, according to Jarmila Kukalova-Peck (pers. 
com.), there is evidence that Embiidina and what she 
regards as its sister order, Plecoptera, evolved from a 
common early Carboniferous ancestor. Tertiary em- 
biid fossils closely resemble extant species — all with 
the same adaptations easing backward, predator-es- 
cape movement in narrow, silk galleries. 

This rather "textbookish" treatment of integumen- 
tal anatomy emphasizes these order-defining struc- 
tures. Especially important are the peculiar wings of 
males. To attain ultimate freedom of movement in 
galleries, females of all species, by the process of 
neoteny, became apterous (Fig. A). 

Much attention is also given to evolution and di- 
versity of the abdominal terminalia of males which 
are of prime importance in defining species and high- 
er taxonomic categories. Such diversity, as well as 
other anatomical features, are profusely illustrated. 
Additional information relating to anatomy will ap- 
pear in EMBIA, Part 2, "A review of the biology of 
Embiidina." 

Relationships 

Based on a phylogenetic assessment of embiid 
wing venation and articulation, using specimens pro- 
vided by me, Kukalova-Peck has concluded that em- 
biids comprise a very old order derived from 
generalized plecopteroid stock and that this stock, 
represented by Upper Carboniferous fossils, must 
have served as a stem group for Plecoptera and Em- 
biidina. Incidentally, she regards plecopteroids as 
the sister group of all other Neoptera. 

Earlier, Hennig (1981) speculated that ancestors 
of Embiidina diverged from the paurometabolan line 
about the beginning of the Carboniferous period. He 
believed that the earliest division of Paurometabola 
was Embiidina and Orthopteromorpha with the lat- 
ter dividing into Blattopteriformia and Orthopteroi- 
dea. Thus, he concluded that embiids have an isolated 
systematic position and high categorical rank. 

Rahle (1970) decided that there is a sister-group 
relationship between Embiidina and Phasmida. Mat- 
suda (1970) believed that Embiidina and Phasmida 
are apomorphic sister groups of Plecoptera (Fig. C). 



Kristensen (1975), while reviewing the phytogeny of 
all hexapod orders and devoting considerable atten- 
tion to Timema (Fig. D), a generahzed phasmid, con- 
cluded that determination of the relationship of 
Embiidina is inconclusive. He believed that further 
comparative study of the thorax of alate male embiids 
(Fig. B) and their genitalia might aid determination of 
the relationship of these orders. Storozhenko ( 1 997) 
placed Embiidina in the superorder Plecopterodea with 
orders Plecoptera, Grylloblattida (Fig. E), Dermaptera, 
and the extinct order Protelytroptera. Phasmids were 
placed with Orthoptera in the superorder Orthopterodea. 
The most thorough treatment of orthopteroid relation- 
ships is that of Flook and Rowell (1998). They decid- 
ed that Embiidina is a sister taxon to a clade consisting 
of Phasmida and Orthoptera. 

Origin 

Embiid ancestors may have had the following 
characteristics: (1) A prognathus head with chewing 
mouthparts and long, filiform antennae, as in extant 
embiids. (2) Legs moderately long, as in Plecoptera; 
tarsi five-segmented. (3) Wings present in both sex- 
es, subequal in size and shape except for a relatively 
large anojugal area in hind wings; veins uniformly 
sclerotized, none developed as blood sinuses. (4) Male 
abdominal terminalia symmetrical, cerci more than two 
segmented. (5) Habitat and food (as today), weath- 
ered outer bark and dead leaf litter. (6) Defensive 
biting weak; repugnant secretions absent. (7) Distri- 
bution, equatorial portions of Laurasia, perhaps dur- 
ing the Carboniferous period. 

The ancestors probably ventured out of crevices 
and other protective places to feed and retreated into 
them to avoid predators and other adversities. It is 
likely that short hind leg adaptations for rapid back- 
ward movement developed during the order's early 
evolution. Such predator avoidance would have been 
greatly improved if feeding forays could have oc- 
curred beneath a silken cover or within galleries. 
Such an unUkely adaptation was yet to evolve. 

During the Carboniferous period there must have 
existed embiid stem-group species with the above- 
described characteristics and habits but, by chance, 
only in one of them did mutations appear which en- 
abled individuals to spin silken coverways using high- 
ly efficient "tools" — the basitarsi of their forelegs. 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



One may well question, as with many other "thresh- 
old innovations," how the initial mutation could have 
been effective. It may be that a single mutation might 
have simultaneously modified each of the many 
secretory pores on the plantar surface of the basal 
segment of the foretarsi. If this had happened, the 
quantity of silk immediately might have been suffi- 
cient to produce at least crude predator-bamers ex- 
tending outward from crevice retreats, or for crevice 
coverage during vulnerable periods such as ecdysis. 
or brood development. 

Offspring of the mutant individuals could have 
survived to reproduce in greater numbers, especial- 
ly, as is likely, the brood had remained together and 
mated incestuously within their parent's "nest." Over 
time, through natural selection, the secretory cells 
became multiple, elaborate internal glands while the 
derm evaginated to produce associated tubular, seta- 
like ejectors (Figs. 19. 20). This now-universal en- 
semble has enabled all instars of all species of the 
order to extend their feeding range within tubular gal- 
leries, often enhanced in bark inhabiting species by 
an additional sheet-like covering. 

The tubular form of the galleries, sized to fit the 
maker, is responsible for specializations which de- 
fine the order, as follows: (1) A linear, supple body 
with short legs. (2) Prognathism. (3) Greatly en- 
larged hind tibial flexor muscles motivating rapid 
backward movement in galleries. (4) Reduction of 
wing-snag by flexibility resulting from desclerotiza- 
tion of most wing veins. (5) Compensating ability 
to stiffen wings for flight by blood pressure in sinus 
veins — especially in the anterior radius (RBS). (6) 
Universal apterism of females, and males of some 
species, through neoteny — the ultimate adaptation for 
improving backward movement within galleries. 

Fossil record 

The few fossils which unquestionably represent 
Embiidina are modem forms from Tertiary forma- 
tions and. therefore, useless for tracing the origin and 
early evolution of the order. Fortunately, however, 
much can be learned fi-om extant, plesiomorphic spe- 
cies such as Amazonia's Clorhoda nobilis (Gerstaek- 
er) and C. longicaiida Ross. Also, even without 
fossils, one can postulate a great age for the order 
because it must have fully evolved its order-defining 
features, its major higher taxa, and wide distribution 
on Pangaea long before its fragmentation. The or- 
der's specializations are much too complex to have 
evolved de novo on each separated tectonic plate. 




Figure C. Paired stoneflies (Plecoptera). Acroneuria 
califoniica (Banks) from northern California. Although 
there are vast anatomical and biological differences between 
modem Plecoptera and Embiidina, it is believed that the 
two orders diverged from the same stock during the Car- 
boniferous period. The copulatory positions are very simi- 
lar but in Plecoptera there are no adaptations for reverse or 
rapid movement. 




Figure D. Adult male of Timema califomica Scudder 
(Phasmida), from central California, on foliage of live oak. 
Its Paleozoic ancestors have been considered possible rela- 
tives of those of Embiidina. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



r 




Figure E. Adult female of Cniloblatta cainpadeifonnis 
occidentalis Silvestri (Grylloblattida) from Mt. Baker, 
Washington (1280 m). Kukalova-Peck (pers. com.) on the 
basis of many characters concluded that grylloblattids be- 
long to the blattoid line. 



piece of very hard, light tan rock. Anatomical de- 
tails are not as clear as represented in Martynova's 
drawings (Fig. 2) and I can see no evidence that these 
fossils represent Embiidina. The head and antennae 
of Sheimia are embiid-like but the caudal cranial 
margin appears to be more transverse, instead of 
rounded, as in embiids. The apparently quadrate pro- 
thorax of Sheimia is embiid-like but this character 
isn"t peculiar to embiids. The pterothorax is well 
sclerotized and broad, not unlike that of embiids, but 
the scuta, figured as triangular by Martynova, are not 
clearly so in the type specimen. The meso- and met- 
athorax, however, seem to be nearly equal in size 
and this embiid feature is associated with wide- 
spaced, nearly equal (homonomous) wings. 



Pretertiary fossils 

Zeuner (1936), treating his Gennanoprisca zim- 
mennanni, a new species from a European Permian 
formation, concluded that Embiidina were derived 
from the family Lemmatophoridae of the Protoper- 
laria. These fossils belong to the plecopteroid stem 
group which possess similar, generalized wing ve- 
nation and attachment. No pretertiary fossil shows 
blood sinus wing veins which characterize embiids 
but, of course, much of the order's evolution must 
predate the appearance of sinus veins. 

In 1937 Tillyard established the new suborder 
Protembiaria (he used Embiaria as the ordinal name) 
and family Protembiidae to include his Protembia 
penniana. the type specimen of which was found in 
Elmo deposits. Lower Permian of Kansas. Shortly 
thereafter, Zalessky (1937) named two additional 
species in his new genus, Tillyardembia. from relat- 
ed Russian Permian deposits. Later, Carpenter 
(1950), after re-examination of Kansas specimens, 
decided that they do not represent Embiidina, and 
assigned Protembia to the cumulative order Protor- 
thoptera. His figure of P. penniana indicates that it 
was a rather long-legged insect without any of the 
specializations characterizing Embiidina. 

In 1958 Martynova proposed a new embiid sub- 
order, Sheimiodea, family Sheimiidae and species, 
Sheimia sojanensis (Figs. 1, 2), based on a winged 
fossil from the Upper Permian of Russia. Through 
the kindness of Dr. Frank M. Carpenter I was able to 
closely examine this specimen which, with two ap- 
parently conspecific individuals, is preserved in a 










FlGURE I . Holotype of Sheimia sojanensis Martynova, 
Upper Permian of Russia. 

..J 




Figure 2. Martynova's representation of the holotype of 
Sheimia sojanensis (Redrawn by Ross). 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



Martynova's drawing (Fig. 2) indicates that the 
femora of the foreleg are greatly enlarged (an embiid 
feature), but it is not certain that the faint impres- 
sions she regarded as femora are indeed femora. The 
wings are quite embiid-like but broader than typical. 
The venation, although delineated as definite in 
Martynova's drawing, is obscure, yet there is a slight 
indication in the specimen of the radius-border lines 
which are peculiar to embiids. If indeed these were 
present, this would be a strong basis for assigning 
Sheimia to Embiidina, or a related extinct order. 

The fossil's abdomen is broad, perhaps due to 
pressure of fossilization. and appears to be weakly 
sclerotized with a darker apex. Unlike embiids, this 
apex tapers to an acute point and Martinova's draw- 
ing indicates an ovipositor-like internal (?) structure, 
but this could simply be a disconnected basal por- 
tion of a filament-Iike cercus. She also indicated what 
appears to be fragments of multisegmented cerci in 
the vicinity of the abdomen's apex. 

After serious consideration I have concluded that 
Sheimia is not an embiid. Carpenter (in litt, 1965) 
drew the same conclusion and later (1976) placed 
Sheimiidae and four other "problem" families in 
"Families of uncertain ordinal position." However, 
there is a strong possibility that it belongs to the 
embiid stem group. 

Tertiary fossils 

Perhaps the oldest Tertiary embiid fossil, Bunnitem- 
bia venosa Cockerell ( 1 9 1 9) is a very small male (body 
length 4.4 mm) preserved in amber (Burmite), pos- 
sessed by the Paleontology Department, British Natu- 
ral History Museum. Some workers, e.g., D. A. 
Grimaldi (pers. com.), consider Burmite to be clearly 
Eocene but others, e.g., A. P. Rasnitsyn (pers. com.), 
believe it to be late Cretaceous. The fossil, definitely 
an embiid, was treated in detail by Davis (1939). I 
briefly examined the specimen, didn't have time to 
study it in detail, but I have decided to assign it to a 
new family. Contrary to expectations, the species is 
highly apomorphic. 

Several embiid specimens have been encountered 
in Baltic Amber (Eocene?) and, fortunately, two of 
these are adult males clearly exhibiting the terminalia. 
Thanks to the Geologischen Staatsinstitut in Ham- 
burg, I have studied these specimens and suspect that 
they are the cotypes of Embia antiqua Pictet (1854). 
I assigned the species (Fig. 36) to a new genus, 
Electroembia Ross (1956), and concluded that it is 



more closely related to Rhagadochir Enderlein and 
related genera of equatorial Africa than to Embia. 
Apterism and melanism of these males suggests that 
the climate during the period of entrapment in amber 
was characterized by a long, dry season. In the Med- 
iteiTanean region there are several species of Embia 
and Haploembia with black, apterous males with gen- 
eral appearance almost identical to E. antiqua and 
can be distinguished only by examination of the ab- 
dominal terminalia and hind tarsi. 

Two small, alate embiids, collected in Chiapas 
Amber, Mexico, are new species of the large modem 
genus Oligembia Davis, family Teratembiidae, of 
which there are many species in the American trop- 
ics. Chiapas Amber has been "positively dated as 
Oligocene and Miocene with possibility in some ar- 
eas of 'Eocene epoch' " (Hurd, Smith and Usinger, 
1958). 

More recently embiids have been collected in Oli- 
gocene amber mines of northern Hispaniola (Domin- 
ican Republic). Most of these fossils, some in my 
collection, are winged males of the genus Oligembia. 
One of these in the American Museum's collection 
became the holotype of Oligembia vetusta Szumik 
(1994). Other Dominican Amber embiids belong to 
two new genera in Anisembiidae, one of which is 
represented today in Hispaniola, southern Mexico and 
Central America. 

Embiids preserved in East African copal, such as 
Oligotoma westwoodi Hagen, represent modern spe- 
cies of the family Teratembiidae. 

The Eocene-Oligocene alate male fossil, Embia 
florissantensis Cockerell, found in Florissant (Colo- 
rado) volcanic shales, a large alate male of the fami- 
ly Embiidae, has been assigned to a new genus, 
Lithembia Ross (1984). The all-important charac- 
ters of the male terminalia are indistinguishable. 

According to A. P. Rasnitsyn (pers. com.), of 
Moscow, the Miocene shale fossil Clothonopsis mi- 
ocenica Hong and Wang ( 1993) isn't an embiid but 
has proved to be a bibionid fly. perhaps genus Plecia. 



Integumental Anatomy 

"Anatomy is what you see with your 

eyes, morphology is what you think you 

see with your mind." 

R. E. Snodgrass 



It follows that the appropriate title of this 

work should be the anatomy and 

morphology of the Embiidina. 



Introduction 

For this simple exposition of the basic integu- 
mental anatomy of Embiidina, a representative spe- 
cies of the order, an OUgotoma. was selected because 
three species of this genus have been widely distrib- 
uted in commerce and, because adult males are at- 
tracted to light, they are most frequently collected 
and available for comparative study. 

Adult females throughout the order are neotenic 
and thus, except for their paragenital stemites, inter- 
nal genitalia, size and coloration, are similar to early 
stage nymphs. Probably due to advantageous endo- 
crinal controls curtailing development, most features 
of adult females are those of an early instar nymph 
before even a trace of developing wings appear (Fig. 
A). As will be discussed later, intermediate stages of 
neoteny have evolved in adult males of many spe- 
cies, as is shown by all ontogenetic stages of wing 
development from complete aptery (Figs. 34, 35, 36) 
through various stages of wing pad growth, brachyp- 
tery and microptery. Fully formed wings, however, 
are the most apparent characteristic of adult males of 
the majority of species (Figs. B, 53). The alate con- 
dition is associated with a full compliment of adult 
embiid anatomical features, many of which certain- 
ly must also have been possessed by adult females 
during early evolution of the order. 

Because of the high degree of specialization for 
silk gallery life exhibited by adult females, it is safe 
to assume that such features evolved not especially 
to foster adult male survival but instead to secure and 
prolong the lives of the more essential adult females. 
Being short-lived and needing only to quickly dis- 
perse and inseminate females, males faced less se- 
lective pressure toward perfect adaptation to gallery 



life and consequently are most likely to retain struc- 
tures, such as wings, which are no longer of critical 
importance in survival. 

Among other plesiomorphic characters, adult 
males also possess structures which aid location of 
mates and insure successful copulation. Because of 
their more complex anatomy, and consequent great- 
er value in systematics, I will devote most attention 
to the anatomy of adult males. 

Methods 

To fully expose sclerites which are partially or 
entirely concealed in integumental folds, almost all 
studied specimens were macerated in heated 10% 
potassium hydroxide (KOH) and cleared of body con- 
tents. Whenever possible drawings were made of 
specimens positioned in alcohol. At times acid fuch- 
sin stain was used to intensify details. Distortion of 
position of sclerites is inevitable in KOH-treated spec- 
imens and this must be considered during interpreta- 
tion of certain drawings. Unless otherwise credited, 
all of these were made by me, based on camera luci- 
da and photographic outlines. I adopted the style of 
R. E. Snodgrass (1935) with membranous areas stip- 
pled and degree of sclerotization represented by thick- 
ness of lines and some shading. Usually setae are 
omitted. The figures are intended to be self-explana- 
tory and the text only a guide and a means of calling 
attention to the functional or phylogenetic significance 
of structures. Unless otherwise indicated, they are 
based on specimens of OUgotoma nigra Hagen. 

Knowledge of musculature is vital to any study 
of integumental anatomy, but I haven't had the time 
nor background to make such a study. Therefore. I 
asked Edward L Smith to at least investigate the mus- 
culature of the abdominal terminalia of adult males 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



and some females. Some of this research has been 
apphed by me in the present paper. Perhaps his com- 
plete research will be published separately. 

I am indebted to Christine A. Nalepa, a cockroach 
specialist, for useful suggestions concerning my treat- 
ment of embiid wings, their reduction and loss. 

While concurrently producing Part 2 of EMBIA 
which reviews the biology of Embiidina. I became 
increasingly awai-e of the difficulty of separating be- 
havior from anatomy — especially in the case of em- 
biids where there is such a close relationship between 
structure and function, as is most apparent in my treat- 
ment of embiid wings. Consequently, some infor- 
mation in this paper will be repeated in my review of 
embiid biology. 

Head 

The head of embiids is orthopteroid in type, basi- 
cally similar in all stages of all species (Fig. 3). Due 



to the universal vegetative diet of nymphs and adult 
females, there has been no great need for head spe- 
cialization and diversification. Only in adult males 
are there significant cranial and mouthpart special- 
izations and these relate to the habit of males of many, 
if not all species, to grasp the female's head with the 
mandibles prior to. or during, copulation. Except in 
highly neotenic males, guts of adult males are emp- 
ty and thus their mandibles aren"t used for eating, 
instead being used only as mating claspers and for 
cutting an entrance into a gallery occupied by a re- 
ceptive female. 

Non-neotenic males also have varied cranial and 
eye structure as well as longer more sensory anten- 
nae (Figs. 4, 34). Such characters probably are relat- 
ed to a male's more frequent movement outside of 
galleries to disperse and locate a mate. Due to neo- 
teny, adult females retain the nymphoid head anato- 
my of early instars of both sexes (Fig. 3C, D) but, as 




B male ventral D female ventral 

Figure 3. General structure of head of .A. B adult male and C. D female of OUgotoma nigra Hagen. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 




Figure 4. Diversity of heads of adult males (labrum not figured in some species). A. Embia. B. New genus 
Embiidae. C. Enveja. D and F new genera, Anisembiidae. E. Pelorembia tumidiceps Ross. Anisembiidae, G. 
and H new genera in a new family. I. Oligembia n. sp., Teratembiidae. J. Australembia. Australembiidae. K. 
Metoligotoma, Australembiidae. L. New genus and species, Oligotomidae. 



10 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



will be later discussed, males of some species also 
have similar nymphoid heads due to neoteny. How- 
ever, at least in one exceptional case, a species of 
Metoligotoma Davis, adult males with completely 
nymphoid heads appear as intraspecific variants in 
populations which include males having an adult-type 
head, normal for the species. 

The head of Oligotoma nigra Hagen. as in all oth- 
er species of the order, is strongly prognathous, an- 
gled downward at about 35° (Fig. 13). Obviously, 
prognathism, an adaptation easing movement in nar- 
row galleries, causes eyes, antennae and mouthpaits 
to have a functional forward position. There is also 
obsolescence of most cranial sutures. 

A feature apparently associated with prognathism 
is the universally-present, broad sclerotic bridge be- 
tween the posterior tentorial pits and the occipital 
foramen (Fig. 3B-D). In many species this bridge, 
here termed "ventral bridge," is so strongly devel- 
oped that all traces of its origin are lacking. In males 
of Oligotoma. however, and more so in those of cer- 
tain other genera, there is evidence that much of this 
bridge is due to postoccipital sclerotization. The pos- 
sibility that the medial portion represents a true gula 
(a posterior extension of the submentum) was con- 
sidered. However, in females and nymphs the sub- 
mentum is clearly delimited basally by a transverse 
suture between the posterior tentorial pits and thus 
the bridge is not continuous with the submentum. A 
sclerotic bridge behind the posterior tentorial pits fre- 
quently occurs in other prognathous insects but in 
such cases a true gula is present. To my knowledge, 
the occurrence of a non-gular bridge in this position 
does not occur in any other order. 

Sutures related to original head segmentation are 
practically obliterated on the venter (posterior por- 
tion) of the head. The post-occipital sulcus can be 
traced, however, by lines on either side of the occip- 
ital foramen and these diminish anteriorly. In males 
there is a slight indication that they parallel the ex- 
tremities of the mesal extension of the posterior cra- 
nial walls on the ventral bridge. They might therefore 
be postoccipital. 

The occipital foramen (Fig. 3B-D) is triangular 
and its margins, although thickened, are not especially 
modified for contact with the first cersical sclerites 
midway on each side. The lateral margins are more 
thickened posteriorly and gradually form a small in- 
ternal ridge extending around the posterior end of the 
head to the dorsal surface where it diminishes. This 



line coincides with the outer margin of the long later- 
al maculation of the pattern visible in the young and 
adults of certain species. 

In adult males, the dorsal (frontal) surface of the 
cranium usually lacks sutures (Fig. 3A), but in fe- 
males, lines of weakness ("coronal sutures") serving 
ecdysis, appear as very faint unpigmented lines ex- 
tending forward to the middle of the dorsum of the 
head where they fork as two broadly divergent, equal- 
ly-faint lines (Fig. 3C). These terminate just before 
reaching the bases of the antennae. 

Both dorsally and ventrally. the cranium often has 
a characteristic pattern which appears to be quite 
uniform throughout the order and is visible in 
nymphs, and adults of some species, as lighter pig- 
mented areas. Melanization of most adults generally 
completely obscures these paler areas, but close ex- 
amination reveals that they can still be traced as ar- 
eas lacking setae and reticulate in sculpture. This 
pattern correlates with interior attachment of bundles 
of mandibular muscles. Another frequent pattern is 
a diffused, often golden, transverse "cloud" from eye 
to eye on an otherwise dark head. This overlays the 
brain and may have some special function, one per- 
haps related to light perception, or to mating. 

The anterior tentorial pits open dorsally as trans- 
verse, short slits situated in the epistomal sulcus just 
behind the anterior articulation of the mandible. The 
epistomal sulcus is represented laterally only, but 
serves to distinguish the clypeus from the frons. The 
anteclypeus (Fig. 3A. C) is entirely membranous and 
limited anteriorly by a transverse fold. 

The antennal socket is surrounded by a basal 
flange delimited by an incomplete antennal suture. 
The space between the eye and the mouth cavity is 
greatly reduced and its sutures are obsolete. 

The tentorium of the male is here figured from 
the anterior and posterior aspects (Fig. 5). The ante- 
rior and posterior arms unite medially and form a 
thin, quadrate corporotentorium. The anterior arms 
bear small dorsal branches which appear to be vesti- 
gial. The posterior arms are dilated and join the very 
large crassa, or hypostoma, and the inwardly-slant- 
ing lateral flanges of the submentum. In males these 
lateral flanges are heavily sclerotized, whereas in 
females they are submembranous. The general struc- 
ture of the tentorium appears to be constant through- 
out the order. The posterior tentorial pits are situated 
on either side of the base of the submentum. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



11 



anterior tentorial 



ontennol socket 



anterior orticulatio 
of mondible 




posterior orticulatio 
of mandible 



A cepholic aspect 



anterior tentorial 




B caudal aspect 

Figure 5. Tentorial structure of adult male of Oligo- 
toma nigra Hageti. 



Eyes 

Paired compound eyes are the only organs of sight 
or light perception, ocelli being absent in all species 
of the order. The eyes of adult males of Oligotoma 
nigra (Fig. 6), as in those of many other non-neoten- 
ic species, are relatively large, inflated, reniform and 
composed of numerous, large, convex facets. The 
greater development of eyes in males usually corre- 
lates with the presence of wings and thus activity 
outside of galleries. Some alate males, however, have 
small eyes. Very large, inflated eyes with prominent 
facets, occur in males of many pale species which 
disperse nocturnally. This combination of pale or 
somber body coloration and large eyes is, of course, 
a condition found in many nocturnal insects, as well 
as in nocturnal vertebrates. 

In females and nymphs throughout the order the 
eyes are very small and have relatively few facets. 
Such a condition in adult females is nymphoid and is 
associated with almost complete confinement in silk 
galleries where activities are probably guided more 
by touch than sight. Males of well-pigmented spe- 
cies, especially those which disperse diurnally, may 
have smaller eyes often with dark pigment in the facet 
interstices. In a few species the head is nearly hol- 
optic; the space between the eyes being very narrow 
and the post-ocular cranial bulk greatly reduced (Fig. 
4B). 




Figure 6. Scanning Electron Micrograph photo of eye of 
adult male of Oligotoma nigra. 



Antennae 

The antennae of both sexes are basically similar 
throughout the order. They are annulated and fili- 
form, thus similar to those of Grylloblatta and nu- 
merous other orthopteroid insects. 

An antenna consists of a scape, a pedicel and 
many flagellar segments (flagellomeres). Few spec- 
imens can be found with a complete number because 
terminal segments frequently are broken or bitten off 
by other embiids. However, these may be partially 
regenerated if lost during an early nymphal instar. 
In general, the number of flagellar segments seems 
to coiTelate with body size; the lai'gest number, about 
32, is present in adults of large species and the small- 
est number, as few as 1 1 , is found in tiny species of 
the family Teratembiidae. Adults of average size, 
such as those of Oligotoma spp., tend to have about 
1 9 flagellar segments. Antennae of the first instar of 
all examined species are nine-segmented, while those 
of the second are twelve-segmented. This increase 
is accomplished by the basal flagellar segment di- 
viding into three. 

Antennae of adult females are nymphoid and sim- 
ilar throughout the order. However, non-nymphoid 



12 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



adult males show consistent intergeneric and inter- 
specific variation useful in classification. Such char- 
acters involve segment proportions, type of vesture 
and color. In many genera, and in some congeneric 
species, the apical segments may be abruptly and con- 
sistently white in both males and females. In many 
genera, e.g., Archembia Ross, females have white- 
tipped antennae while males of most species of this 
genus have uniformly brown segments. Great an- 
tennal length in large males is due in part to increased 
segment-number, but also to segment-elongation. 

As is to be expected in insects that venture into 
open environments to locate a mate, antennae of 
males have the greatest sensory function. Antennae 
of males of two species were studied by Slifer and 
Sekhon ( 1973) with specimens provided by me. They 
reported at least five types of sense organs on flagel- 
lar segments of Ptilocerembia sp. which, incidental- 
ly, has the most hirsute antennae found in the order. 

They concluded that the large sense organs are 
"thick-walled chemoreceptors" (or contact chemore- 
ceptors). Each has about five neurones at the base 
and extend their dendrites to the tip of the hair with- 
in a long cuticular sheath, at which point they are 
exposed to air. 

Mouthparts 

The mouthparts are typically orthopteroid and 
need not be described in detail. The labrum (Fig. 7) 
is a simple, semicircular flap, slightly apically 
emarginated. and membranous medially along its an- 
terior margin, and clothed with setae. The ventral 
surface, or epipharynx. is entirely membranous and 
characterized by two widely spaced, nearly parallel 
rows of inwardly-directed short setae. The surface 
between these rows has a fine reticulated structure 
which continues on the dorsal lining of the esopha- 
gus. Well-developed tormae are located near each 
basal angle of the labrum, and extend well within the 
anteclypeus. The hypopharynx (Fig. 8) is large, sim- 
ilar to that found in other orthopteroids. Its dorsal 
(anterior) surface lies against the epipharynx and is 
clothed with dense, scale-like setae. 

Mandibles of nymphs and adult females (Fig. D- 
F) are stout, asymmetrical, and have large, 
multidentate grinding surfaces weU adapted for chew- 
ing coarse food. The points of dorsal (anterior) and 
ventral (posterior) articulation are located medially 
in nymphs and females, whereas in adult males they 
are located near the caudal anale of each mandible. 



zygu 




Figure 7. Labrum of female Oligotoma nigra. A. dorsal 
B. ventral 



The mandibles of adult males (Fig. 9 A, C) gener- 
ally greatly differ from those of females, as well as 
those of their nymphs the mandibles of which are 
identical to those of adult females. These differenc- 
es are quite constant within a taxonomic group and 
thus important in systematics. Unlike those of fe- 
males, mandibles of males are often elongate, rather 
flattened and without grinding surfaces. Their teeth 
are few in number and located apically, usually with 
three on the left mandible and two on the right. At 
times these apical, or incisor teeth, may be fused and/ 
or curled ventrad (as in the genus Embia Latr.). In 
all Anisembiidae, and certain other taxa, there is fu- 
sion of the subapical teeth with the apicals. 

In males of many species the inner face of each 
mandible has a prominent, acute point sub-basally. 
This, which may be the proxadental cusp, separates 
the incisor area from the molar area, which is often 
greatly projected mesad. Between the incisor teeth 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBllDINA. PART 1 



13 




Figure 8. Lateral aspect of hypopharynx. 



and the proxadental cusp the surface may be smoothly, 
inwardly arcuate but often it is produced as an 
obtusely angulated point, here termed medial flange. 
The outer, sub-basal angle of the mandible is 
occasionally extensively produced. The diverse forms 
of male mandibles are illustrated in Figure 4, as well 
as in my publications dealing with embiid systematics. 

Mandibles of adult males usually differ so great- 
ly from a type adapted for a herbivorous diet that 
some workers erroneously concluded that they must 
be used for predation. Actually, the peculiar mandi- 
bles of many species function as claspers to control 
the head of females during copulation. Only in a 
few species with highly neotenic males are the man- 
dibles ever used for mastication of food because 
males normally stop eating during the penultimate 
instar. It is possible that the mandibles of males may 
be effective in warding off small predators and rival 
males. They are also useful for snipping an entrance 
into silk galleries likely to contain a prospective mate. 

Females and nymphs regularly use their mandi- 
bles to pick up and place fecal pellets outside of gal- 
lery walls. They are also used by ovipositing females 
of some species to pulverize feces and habitat mate- 
rials for inclusion in a paste deposited around eggs. 
Such pulverizations may also be deposited on the 
outer surface of a silk labyrinth to enhance or per- 
haps obscure the protective cover. The mandibles 
may also be used to gnaw burrows into the habitat — 
soil, bark, dead twigs, etc. 



The maxilla (Fig. 10) have no peculiar features 
and are similar in both sexes throughout the order. 
The maxillary palpi of adult males oi Australembia 
Ross, however, are exceptionally large and may as- 
sist mandibular clasping during copulation. The base 
of the cardo articulates well within the head. 

The labium (Fig. 11), although generalized, ex- 
hibits interesting modifications. The glossae once 
were believed to be supplementary spinning organs 
(Enderlein, 1912). This is not the case, however, and 
no trace of labial gland openings have been found 
associated with them. The mentum is greatly reduced 
and is almost obsolete in adult males of many spe- 
cies. In neotenic, apterous adult males and females, 
as well as in nymphs, it is a small triangular sclerite. 

The submentum is a well-developed broad plate, 
weakly sclerotized and uniform in females and young, 
but in adult males of OUgotoma Hagen (Fig. UA) 
and those of many other genera, it is heavily sclero- 
tized with the anterior and lateral margins folded 
inward, or inflexed with lateral flanges and fused ba- 
sally to the posterior tentorial arms as described be- 
fore. In some genera, such as Aiitipaluria Enderlein, 
the submentum of the male is deeply bifoveate and 
these depressions may be a source of glandular se- 
cretions associated with mating. In Archembia Ross 
the surface may be flat and from which tiny pores 
produce a secretion that becomes a white coagulant 
around setae in alcohol-preserved specimens. Males 
of some species of a new southeast Asian genus have 
a peculiar, rugose, conate submentum but in other 
species of the genus the structure is normal. The pos- 
terior tentorial pits are located at the proximal angles 
of the submentum and are separated from the occip- 
ital foramen by the extensive ventral bridge. 

Cervix 

The large size of the cervical sclerites (Fig. 12 B, 
C) appears to be related to vigorous head activity. 
For example, strong musculature is involved in head- 
pushing to shape silk galleries as they are spun and 
adult males of many species need cervical strength 
to pull a female's head to the right in a mandibular 
grip prior to copulation. Adult males are also able to 
turn their heads at surprisingly extreme angles as, in 
a mantid-like manner, they follow the movements of 
an observer, most often by males resting on walls 
under a light. 



14 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




condyle 
(pre artis) 



adductor muscle 

attachment 

(rectotendon ) 



outer 
angle 



,- , / I extenso 
■ (pre artis) i attachment 
post artis 



condyle 
(pre artis) 



incisor teeth 

denies ^ 




adductor , extenso 

muscle attachment ; (gndon 

" - condyle (pre artis) attachment 



post artis 
Figure 9. Mandibles of adult male (A-C) and adult female (D-Fj 



The first lateral cervical sclerites contact subme- 
dial points on the sides of the occipital foramen but 
tJiese points are not especially modified to receive 
such contact. The outer edge of each cervical scler- 
ite is thickened and its anterior apex bears a small 
hair plate. In the head's normal position, about 35° 
downward from the horizontal axis (Fig. 13), the sec- 
ond lateral cervical sclerites are almost perpendicu- 
lar to the longitudinal axis of the first. This angle 
becomes obtuse as the head projects forward and is 
thus more horizontal. Each second lateral cervical 
sclerite articulates with a small, medio-ventral pro- 
jection of the adjacent pronotal flange. 



In the ventral neck membrane, immediately be- 
fore the prostemum, are located small sclerites; the 
anterior of which is smaller than the posterior (Fig. 
12B). Crampton (1926) designated the first as an 
interstemite and the second as the presternum (an an- 
terior detachment of the prostemum). In the present 
work these are simply designated anterior and poste- 
rior ventral cervical sclerites. Ahead of these, between 
apices of the first cervical sclerites, are two membra- 
nous domes which are densely setose in Clothoda 
Enderlein. Similar setose domes are found in mem- 
branes lateral to these apices. These are termed latero- 
ventral cervical sclerites. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



15 



lacinia 



cranial 
articulation 



FiGLrRE 10. Maxilla, adult male. 





Figure 11. Labium, adult male (A), adult female (Bj. 



There is no great intra-ordinal variation in the cer- 
vix except in males of certain genera. In Enveja 
Navas the lateral cervical sclerites are especially 
broadened and this seems to be associated with the 



need for greater anchorage of large muscles moving 
an exceptionally large head. In this case the second 
lateral cervical sclerites are greatly produced me- 
sad, almost touching the presternal sclerite. 

Prothorax 

The prothorax (Fig. 12A, C), one of the least 
specialized parts of the body, is similar in both 
sexes throughout the order; the prothorax of the 
alate male simply being less robu.st. The generalized 
condition reflects a lack of unusual prothoracic in- 
volvement in embiid behavior. 

The pronotum is unusual, however, in that it is 
not produced laterally in folds and is usually much 
nan'ower than the pterothorax. It is divided across 
the anterior third by a deep furrow which delimits a 
narrow anterior part, the prescutum, and a large con- 
vex posterior part. The latter is divided medially by 
a longitudinal suture probably associated with ecdys- 
is. In males the posterior angles are broadly pro- 
duced caudad as gradually diminishing extensions. 
Medially, the caudal margin is developed as a small 
point. In many cases, especially in nymphs and fe- 
males, the pronotum bears a characteristic pigment- 
ed design which lacks setae and is reticulated, as in 
the head pattern. This design, like that on the crani- 
um's vertex, probably is related to internal muscle 
attachment. In males of many species the prothorax 
often is pale in coloration, or reddish in some diur- 
nally dispersing species. 

The very clearly-defined straight sutures which 
form the lateral margins of the pronotum delimit flap- 
like flanges on either side. These invaginate posteri- 
orly to form two wide-mouthed, pleural apophyseal 
pits, the apophyses of which are stout, conical and 
directed caudad at a 45° angle. Submedially, a small 
ventral lobe is produced against which the base of 
the second cervical sclerite aiticulates. Behind this 
projection a suture extends to the mouth of the apo- 
physeal pit and defines the catepisternite as a dis- 
tinct sclerite somewhat as in Plecoptera. This sclerite 
is slightly convex and produced ventrad in the poste- 
rior half to form the dorsal coxal articulation. The 
trochantin is dorsally separated from the catepister- 
nite by a transverse suture. 

The prothoracic steiTium (Fig. 12B) is a broad, 
quadrate plate, the posterior angles of which are pro- 
duced as short truncate processes. Narrow, lateral 
areas separated by sutures may be subcoxal elements 
fused to the prostemum. Slit-like openings of the 



16 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




sternal 
apophysis 



B 



pronotal flange 
catepisternite 



pronotum pleural apaphysis 




lateral ' posterior ventral 

cervical sclerites cervical sclerite 



sternal opophysis ,' 

post sternum 



Figure 12. Structure of prothorax. adult male. A. Dorsal aspect, B. Ventral aspect, C. Lateral aspect. Oligotoma nigra 
Hagen. 



sternum's apophyseal pits are located on either side 
of the posterior margin. These are transverse and 
develop internally as short, terminally-dilated, flat 
apodemes. The broad area between the inner ends of 
the apophyseal slits represents the point of basal con- 
tact of the very large first poststemum which is fused 
to the prostemum. 



A detailed treatment of the prothorax, including 
its musculature, will be found in Bitsch and Raymond 
(1970). 

Pterothorax of males 

The pterothorax of alate males (Figs. 13-16) ex- 
hibits a number of very interesting features. Unlike 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



17 




Figure 13. Lateral aspect of head and thorax of a male showing angle of head in repose. Somewhat distended as a result 
of KOH maceration. Oligotoina nigra Hagen. 



corresponding somites of females which are round- 
ed and distended by a large, food-filled crop, these 
somites are dorso-ventrally depressed and transverse- 
ly rectangulate in cross-section in alate males. 

Males of Oligotoma nigra have a rather apomor- 
phic type of embiid pterothorax but adequately repre- 
sent general features. Most prominent is the large, 
elongate, triangular mesonotum (scutum 2) which 
abruptly arches downward anteriorly fonning a near- 
ly vertical prescutum terminating in a short phiagma. 
The acrotergite (anterior notal plate) is quite similar 
to that of females and nymphs, being separated from 
the antecostal suture by a membrane. The anterior 
notal wing processes are prominent, strongly devel- 
oped, with deep submembranous emai-ginations be- 
hind them. The posterior notal wing processes, located 
not far behind the former, are nearly obsolete, being 
represented only by the anterior angles of very weak- 
ly-sclerotized lateral flanges of the mesonomm. The 
mesonomm is rather strongly airhed in cross-section 
and the lateral flanges are limited mesad by a low point 
of the sides of this arched portion and hence are di- 
rected upward. In apomorphic families, such as Olig- 
otomidae, Anisembiidae and Teratembiidae, these 
lateral areas are very weak. In plesiomorphic groups, 
such as Clothodidae and Embiidae, they are more 
heavily sclerotized. When wings are in repose, a cer- 
tain amount of inward bending occurs along this area 
and the tendency for its weakness may be an adapta- 
tion for life in narrow galleries, for such bending per- 
mits the wings to rest more directly over the body (Fig. 
B). Otherwise, wing-edge projection beyond the 
body's lateral lines could cause disadvantageous fric- 
tion with adjacent gallery walls. The most peculiar 
feature of the pterothorax is the very great reduction 
of the scutellum which causes the axillary cords to 



almost meet on the mid-dorsal line. To my knowl- 
edge, these are the longest axillary cords to be found 
among insects. 

The acrotergite (anterior notal plate) of the me- 
sothorax is exceptionally large and closely contacts 
the .scutellum and the metathoracic prescutum. It is 
divided medially by a longitudinal membranous cleft. 
Small, elongate sclerites located on either side be- 
neath the axillary cord probably represent isolated 
rudiments of the anterior angles of the acrotergite. 
Essentially, the metanotum has the same structure as 
the mesonotum. It differs chiefly in its shorter pro- 
portions and the fact that its acrotergite forms a bridge 
between the scutellum and the first abdominal ter- 
gum to which it is fused. 

In males the pterothoracic pleura (Fig. 16B) are, 
of course, more extensively developed than in females. 
A peculiar feature is the nearly vertical prealar sumre 
near the anterior end of the episteiTium which delim- 
its a small sclerotic area (prealar sclerite). This area 
may represent a fusion of one of the anterior precox- 
al sclerites, which in females is separated from the 
episteiTium. Another possibility is that it results from 
secondai7 folding. The suture is represented inter- 
nally as a strong ridge directed caudad. The dorsal 
end of the prealar suture forms a process separated 
from the pleural wing process by a membranous area. 
The basalare bridges the gap between these two pro- 
cesses in the fonn of a small, elongate, convex arch 
which is firnily attached at either extremity. The sub- 
alar, a small unmodified scleilte, appears to be simi- 
lar in males throughout the order. 

The sternum of the pterothorax (Fig. 15B) is a 
composite plate resulting from fusion of precoxal 
sclerites (free in females and young) with the basis- 
ternum and furcasternum. This fusion can readilv be 



18 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




spiracular lobe-- 

acrotergite - ■ 

prescutum - - . 

tegula - - _ . 

pteralia \ 

lateral flange ,^ 
axillary cord - 

— episternum 

-pleural suture- 
epimeron 

scutellum 



acrotergite- - 
— prescutum - - 
pteralia."'' 



-episternum - - 

pleural suture- 
- epimeron - - 

acrotergite- - 



^ laterotergite 

abdominal 
spiracle 1 







coxa 3 



A B 

Figure 14. Meso- and metathorax. dorsal aspect of A. Adult female, and B. Adult male. OUgotoma nigra Hagen. 



traced by lines of internal thickening, by the position 
of the sternal apophyseal pits, and often by distribu- 
tion of setae. TTie sternum of a male of OUgotoma 
nigra has been figured alongside that of the female 
(Fig. 15A, B) and a comparison clearly indicates the 
homologies of the various areas of the composite ster- 
num. The subcoxal area. \erN' prominent in the me- 
sostemum. is obsolete in the metastemum of this 
species, but is more evident in more plesiomorphic 
genera, such as Emhia. 

The pterothorax of more plesiomorphic Embiid- 
ina, such as clothodids and embiids. seems to retain 



man) primitive features, e.g., broader and shorter pro- 
portions of the scutum with stronger lateral flanges, 
a broader scutellum. a partially separated prescutum, 
undi\ided acrotergites (notal plates), as well as cer- 
tain features of the sternum. 

Meso- and metathorax of females 

The meso- and metathoracic anatomy of adult fe- 
males of OUgotoma nigra (Figs. 14 A. 15 A. 16A) typ- 
ifies that of adult females and nymphs throughout the 
order, as well as that of fully neotenic (apterous) males. 
Adult females unquestionably once possessed wings. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART I 



19 




■ apophyseal pit 1 



preolor suture --i 
anapleural cleft 

-- preepisternum - 

episternum- 

apophyseal pit 2- 

trochantin ' ' 



"'' ventropleurite ' 
- — poststernum - 

prealar suture - 



furcasternum . 
^' preepisternum -. 

. episternum . 

- -ventropleurite - 

apophyseal pit 3 - ' 
trochontin 



coxa 3 

" abdominal sternum 1 




A B 

Figure 15. Meso- and metathorax. ventral aspect of A. Adult female and B. Adult male. Oligowma nigra Hagen. 



Their present simple juvenile thoracic structure strongly 
contrasts with ancestral complexity retained in the tho- 
rax of alate males. Thoracic simplicity in adult females 
obviously results from neoteny for its sclerotization is 
nearly identical to that of immature stages of females, 
and the second instar of males destined to develop 
wings. Some difficulties in interpretation can perhaps 
be attributed to the fact that most sclerites are rather 
soft and are seldom fused along sutural lines. This 
would appear to make possible distention of the midg- 
ut and more supple body movement — the latter a de- 
cided advantage in gallery life. The apterous condition 



of adult females, and the same tendency in males of 
some species, is undoubtedly an adaptation for life in 
silk galleries, as will be fully treated in the section deal- 
ing with wings. 

The meso- and metathoracic scuta of females are 
simple, large plates, each having a prominent anteri- 
or acrotergite. The narrow transverse sclerite in the 
membrane between the anterior margin of each and 
its acrotergite is probably homologous to the fully 
developed prescutum in alate males. There is no great 
development of phragmata. The posterior margin of 
the metathoracic scutum closely contacts the first 



20 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



prescutum scutum 2 acrotergite prescutum scutum 3 

» 'I ' abdominal abdomrnoi 

acrotergite .^ pleural suture ; epimeron2 '^ \ pleural suture | epimeron3 terga spiracles 

/ \ ,1 2 




prealar ^ '^ — i ^'^i/ ' 






_— ^!X^<1_;^ 


sclerite ' i / ' ' 


/ 




1 ' 


poststernum ' / ^' 






coxa 2 1 


episternum / /' 








preepisternum ' 


ape 


physea 


pit 1 


basisternum 






poststernum 



spiracle /'',',,'' \ \ 

/ \ I \ trochantin 3 ^ \ 

basi sternum , i > l i -^ ^ l j - i 

'^ ' apophyseal pit \ abdominal 



preepisternum 



episternum 
furcQSternum 



\ Sternum 2 
coxa 3 



prescutum pleural suture 
acrotergite \ wing base 'i 



prescutum pleural suture acrotergite abdominal 



tergum spiracles 




spiracle . 



poststernum 



prealar suture i preepisternum i coxa 2 spiracli 



prealar sclerite 



basisternum trochantin 



B 



preepisternum ^ coxa 3 



basisternum 



1^ abdominal 

trochantin sternum 



Figure 16. Meso- and metathorax. lateral aspect of A. Adult female and B. Adult male. 



abdominal tergum without any apparent development 
of an acrotergite of the latter. 

The pleura of both thoracic somites are similar (Fig. 
16A). The pleural suture dorsally contacts the anterior 
angle of the scutum and extends ventro-caudad at an 
angle of about 30" to form the dorsal articulation point 
of each coxa. Coxae are located at points directly in 
line with the posterior angles of the scutum. This dor- 
so-ventral compression, characteristic of all embtids. 
probably is associated with life in narrow galleries for 
it results in a lengthening of the body and a decrease in 
its vertical plane. On either side of each pleural suture 



is a narrow epimeron and a slightly broader epister- 
num. The mesothoracic episternum produces a ven- 
tral trochantin which, in the metathorax, is delimited 
by a deep cleft extending nearly to the pleural suture. 
The apices of the tronchantins constitute the ventral 
articulation of the coxae. 

Between the ventral (anterior) margin of the epis- 
ternum and the basisternum there are large, weakly 
sclerotLzed plates which I believe are pleural in origin. 
In the mesothorax they are rather rhomboid in shape, 
unequal in size and. as in the metathorax, are two in 
number, probably as a result of secondary division. In 



ROSS; EMBIA, BIOSYSTEMATICS OF THE ORDER EMBflDINA. PART I 



21 



the metathorax they ai"e unequal; the anterior sclerite 
(pre-epistemum) being larger and greatly elongated, 
the posterior (ventropleurite) smaller. In an alate 
male's thorax, the ventropleurite is fused to the sides 
of the furcastemum and the pre-epistemum becomes 
the anterior appendix (prealar area) of the adult male 
epistemum. Beneath each coxa there is a small quad- 
rate sclerite surrounded by membrane which, in adult 
males, also becomes a part of the sternal plate. Ten- 
tatively, I regard this series of three sclerites as the 
precoxal arc. 

The sternal region (Fig. 15 A. B) is one of the most 
interesting portions of the thorax of Embiidina. Mat- 
suda's (1960) interpretation of the thoracic sternites 
of insects seems to satisfactorily explain the condi- 
tions present. The development of the sternum in 
the order is one of the best examples of the broad 
type characteristic of orthopteroid orders. This is 
manifested by widely-spaced sternal apophyseal pits 
and a consequently large furcastemum between them, 
as is best exhibited in the mesothorax. 

In the mesothorax of females the sternal apophy- 
seal pits are located in the lateral membranes of the 
sternum and are widely separated by the broad fur- 
castemum which, in combination with a greatly de- 
veloped basistemum, covers most of the venter of 
the mesothorax. The sternal apophyses are very slen- 
der and project only a short distance inward. 

The sternum of the metathorax is similar to that 
of the preceding somite but shorter, with its apophy- 
seal pits very close together. The pits may have "mi- 
grated'" caudad from lateral positions comparable to 
those of the mesosternum. The sternal apophyses 
are much stouter and longer than those of the me- 
sothorax and are strongly directed dorso-laterad. 

The first pair of thoracic spiracles is located in 
the intersomital membrane behind the prothorax on 
prominent, setiferous. lateral lobes and they are ac- 
companied by small sclerites. The second pair of 
spiracles is subventrally located near the anterior an- 
gles of the metathoracic basistemum. They are in- 
conspicuous and do not open on prominent lobes. 

Barlet (1985) has published a more detailed treat- 
ment of the thorax. 

Legs 

The legs of embiids, remarkably similar in all 
taxa, are very short relative to body size as best ex- 
hibited by nymphs, adult females, and neotenic adult 



males, which are more perfectly adapted to gallery 
life than alate, slightly longer-legged, non-neotenic 
adult males. Also universal is the distinct form and 
function of each pair of legs. Most significant are 
the unique foretarsi which produce the peculiar en- 
vironment and thus the order's associated anatomi- 
cal and behavioral adaptations. It is noteworthy that 
all leg features, even the spinning glands, are similar 
in all instars of all species of the order. 

Prothoracic legs 

Silk-spinning is the most important function of 
the forelegs and, accordingly, all of its segments are 
enlarged, well sclerotized and muscled to serve vig- 
orous activity (Fig. 17). The resting position of the 
forelegs is forward, tarsi paralleling the sides of the 
head. During spinning the legs sweep very wide 
arcs — even back and up to the mid-line of the tho- 
rax. This activity includes short strokes in many 
directions combined with outward pushes of the tar- 
si and head to shape galleries. Even teneral individ- 
uals spin soon after ecdysis. Diminutive forelegs 
regenerated after loss during a nymphal instar are 
also capable of spinning. 

The spinning organs (Figs. 18, 19) have been de- 
scribed several times (e.g., Melander, 1902: Rimsky- 
Korsakov, 1914; Mukerji, 1928; bestbyBarth, 1954; 
and Alberti and Storck, 1976). The contention of 
Enderlein ( 1 909. 1912), who probably never observed 
a live embiid, that the silk is produced by labial 
glands, has long been rejected. 

The glands are confined to the greatly enlarged 
basal segment of foretarsus and number at least 150 
per tarsus. Variations according to taxa and instar 
remain to be studied. Each gland consists of a large 
lumen enclosed in an iiTCgular layer of syncytial cells. 
The globular, but often quadrate and in-egular shaped 
glands, are closely appressed to each other They are 
sufficiently large to be visible under low magnifica- 
tion, especially through the pale thin derm of the plan- 
tar surface. The crowded glands resemble seeds in a 
pomegranate fruit. 

Within each lumen there is a chambered corbicu- 
lum (ampulla) with radiating filaments which appar- 
ently direct liquid silk secretions of gland cells into a 
duct (one per gland). Each duct opens distally at the 
tip of a relatively long, hollow filament, here termed 
silk-ejector (ejector of Barth, 1954). They are not 
homologous to setae (Fig. 20A-C). 



22 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 




- trochanter 

B mesothoracic leg 




C tarsus of mesothorocic leg 

Figure 17. A. Prothoracic leg. B. Mesothoracic leg. and 
C. Its tarsus. 




Rgure 18. External aspects of foretarsus of a female. 



According to Earth's excellent cytologicai inves- 
tigation — the first made with the aid of an electron 
microscope — the silk-ejectors and their glands are 
unique organs. As shown in my diagrammed hypoth- 
esis (Fig. 19 A), the glands perhaps evolved from in- 
vaginations of secretory cell pores in the ectodermis. 
Each silk-ejector apparently represents a setae-like, 
evagination of the exocuticle which might first have 
been simply a cuticular rim around a pore opening. 

It is probable that the early ectodermal glands 
were hollow "balls" composed of walled secretory 
cells. The duct, or constriction, leading to each pri- 
mordial silk-ejector must at first have been fully cell- 
lined. The inner walls of these duct cells appear to 
have gradually thickened and fused to become the 
elaborate ducts which now extend from the silk glands 
to the ejectors. Similar ectodermal glands are treat- 
ed and illustrated by Snodgrass (1935: 62, fig. 32). 

Cells of the ectodermal gland and the duct later 
became syncytial. Only a small amount of nucleated 
cytoplasm persists on the outer wall of a duct. The 
ducts are thread-like and wend their way between 
the glands with never more than one duct per gland 
and its ejector. The evolution of such unique, com- 
plex organs is difficult to comprehend unless one con- 
siders the probability that even small mutations must 
have had simultaneous expression on all glands. 
Under such circumstances, even minute, favorable 
modifications could have had almost immediate sig- 
nificance in the survival of the possessor. 

The basal tarsal segment (Fig. 18) is elongate- 
oval. In alate males it is narrower and more cylindri- 
cal in cross-section than in females, nymphs and 
nymphoid apterous males. In females and nymphs 
the segment is almost triangular in cross-section due 
to slant of the inner-dorsal surface. The outer side is 
almost vertical. The inner-dorsal surface is almost 
entirely submembranous, finely reticulate in texture 
and almost lacks setae. A medial, longitudinal de- 
pression which extends the length of the surface grad- 
ually becomes stronger terminally. The area 
immediately inward of this depression often pulsates. 
Dorsal surfaces adjacent to this membrane are darker 
and clothed with moderately long setae of the usual 
type. 

The ventral, or plantar, surface of the basal seg- 
ment (Fig. 20 A) is entirely membranous, pale, often 
pinkish in color, and densely covered with short. 



ROSS: EMBIA, BI0SYSTEN4AT1CS OF THE ORDER EMBIIDINA. PART 1 



23 




Figure 19. A. Hypothetical origin of silk gland. B. Defini- 
tive structure of a single gland and its ejector. C. Glandular 
content of foretarsus of a female. 



acutely-tapered microspicules which may function 
as combs. Silk-ejectors on the outer edge arise with- 
in small, circular areas lacking such microspicules. 
The ejectors are characterized by a lack of basal sock- 
ets and are thin walled, fragile and whitish — condi- 
tions apparently associated with their hollow nature. 
They are variable in form and length but are straight 
except for occasional abrupt curvature of their com- 
plex apices (Fig. 20B, C). 

Along the lateral margins of the plantar surface 
the microspicules are much shorter and the bare spots 
from which the ejectors arise are more conspicuous. 
Ejectors arising from these lateral borders appear to 
be much longer than elsewhere and probably are the 
most important silk-strand ejectors. 

The mid-segment of the foretarsus is a small tri- 
angular pad with a membranous ventral surface 
clothed with both microspicules and silk-ejectors; the 
latter being denser but. of course, less numerous than 
on the basal segment. These ejectors are served by 
ducts arising in glands located within the basal seg- 
ment. The distal segment of the foretarsus is smaller 
than those on the mid- and hindlegs but otherwise 
anatomically similar. During spinning it is elevated 
to prevent its claws from hooking into the webs. 



Nature of embiid silk 

Using specimens provided by me, K. M. Rudall 
of the Department of Biophysics, University of Leeds, 
England, made brief studies of the silk during 1973- 
78 and in letters to me, conveyed the following in- 
formation. 

Double X-ray diffraction pattern shows embiid 
silk to be of the classic Group 1 fibroin silk charac- 
teristic of Bombyx mori and Beria sp. (Cymbidae). 
This silk group isn't found in most other Lepidoptera, 
therefore, its occurrence in an embiid is "most excit- 
ing." He regarded the dermal tarsal glands of embi- 
ids as perhaps simpler secretory systems than the long 
salivary glands of Lepidoptera larvae. The diffrac- 
tion pattern has been interpreted (by universal agree- 
ment) to be due to alternative residues along extended 
polypeptide chains of glycene and alanine. Some of 
the alanine positions are reflected by serine (serine 
is nearly the same size as alanine). Rudall noted that 
the main difference between Bombyx mori silk and 
that of embiids is that the serine to alanine ratio is 
reversed, there being a much greater content of serine. 
The ratios in the embiid sample was 197 glycine to 
40 alanine to 1 30 serine. 

Mesothoracic legs 

The mesothoracic legs (Fig. 17B, C) are the least 
specialized of the three pairs. Their most notable 
feature is a relatively great reduction in size and they 
do not seem to have much locomotor importance in 
or out of galleries. In their normal position they ex- 
tend laterad and span a gallery's interior. They are 
capable of great upward movement and frequently 
the tarsus of one of the pair is able to contact the 
upper surface of a gallery while the body is in a nor- 
mal position. It is possible that these legs aid rota- 
tion of the body within a gallery. 

The tarsi consist of elongated, unmodified seg- 
ments (Fig. 1 7C). The basal segment is evenly clothed 
ventrally with stout setae. The mid-tarsal segment is 
small and has a distal papilla. In adult females and 
nymphs of many species, the ventral membrane of 
this papilla is clothed with very small, basally-directed 
peg-setae which must aid reverse traction on the silk 
substrate. In adult males this surface never is echinu- 
late and perhaps this is another indication of the male's 
poorer adaptation to gallery life. 



24 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure 20. A. Silk-ejectors (white) and "combs" of female Embia (SEM 440x). B. and C. Silk-ejectors and "combs of 
female Embia (SEM 1750x). Ejectors are highly variable in length and distal structure. 



Metathoracic legs 

Formerly, there was some question as to the func- 
tion of the greatly enlarged femora of the hind legs 
(Fig. 21). It was assumed that their enlargeinent 
indicated a saltatorial function. Davis (1936) exam- 
ined the tibial muscles and noted that, unlike saltato- 
rial insects which have large hind tibial extensor or 
levator muscles, the flexor or depressor muscle is 
greatly enlarged and thus accounts for the large size 
of the femora. The extensor, or levator, muscle is 
much reduced and fits into a groove atop the flexor 
(Fig. 21C, D). 

Such tibial musculature obviously facilitates 
backward movement in the galleries — an activity re- 
quiring strong muscles to flex the hind legs. Rapid 
reverse movement in narrow galleries has been a ma- 
jor factor in the evolution of many adaptive charac- 
teristics of the order, notably wing modifications and, 
ultimately, wing elimination in all females as well as 
in males of many species. This will be more fully 
discussed in the section dealing with wings. 

Reverse movement can be rapid, very smooth, 
the body axis remaining straight. In contrast, an em- 
biid's forward movement usually is rather awkward 
and slow but, with stimulation, it can be very fleet 
especially within the galleries. The resting position 
of the hind legs is straight back, closely paralleling 
the sides of the abdomen. In walking, the terminal 
segment of the hind tarsus is elevated and doesn't 



contact the web and, obviously, this avoids a snag- 
ging of the claws in the silk. Thus, substrate contact 
of the tarsus is with the basal and middle segments. 

The basal segment, or basitarsus, probably a com- 
posite of three fused ancestral segments, is slender 
in adult males, stout in nyinphs and adult females 
(Fig. 22). In both sexes the plantar surface bears large, 
irregular, peg-like setae which are more slender in 
males. In some females the setae are directed basad 
in the distal half of the segment. The upper and lat- 
eral surfaces are clothed with long, slender setae 
which, like many others on the legs, have the outer 
curvature finely serrate. Scanning electronic micro- 
scope images show each serration as the apex of one 
of the fibers composing a seta. 

The distal end of the plantar surface of the basi- 
tarsus is always produced as a membranous "blad- 
der." or papilla. Many species, however — often all 
species of a taxon — have a submedial, second papil- 
la on the ventral surface (Fig. 22B). This may repre- 
sent one of the papillae of the three segments which 
fused to form the basitarsus. If so, its presence may 
be plesiomorphic. Two papillae are possessed by spe- 
cies of Clothoda, the order's most plesiomorphic ge- 
nus. However, a second papilla may appear 
sporadically throughout the order as a specific or even 
generic character without any phylogenetic signifi- 
cance. In spite of this, the presence or absence of a 
second papilla is constant in all instars of a species, 
usually so within a genus and thus serves as an im- 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 1 



25 




D cross section 
of 
hind femur 



Figure 21. A. Hind leg of male. B. Hind leg of female, 
C and D. Tibial musculature. 



portant character in systematics. In a few species, 
males have only one papilla, females two. Occasion- 
ally, the second papilla is much reduced, often sim- 
ply a small, unprotruded, clear spot. Such papillae 
appear to be homologous to arolia of the tarsi of grass- 
hoppers and other insects and were so designated by 
Imms (1913). Snodgrass (1935:198) called them tar- 
sal pulvilli. or euplantulae. 



Wings 



Ancestral embiids must have possessed fairly in- 
flexible wings similar in texture to those of most oth- 
er alate insects. However, with increasing dependence 
on quick reverse movement in silk galleries as a prin- 
ciple means of evading predators, the apices of such 
relatively stiff wings must have frequently snagged 
against opposing gallery walls and slowed or prevent- 
ed escape. To overcome such a problem, and to in- 
crease suppleness during U-turns, embiids long ago 
evolved extraordinary wing flexibility. As a result, 
when in repose over the back, the wings of all mod- 



ern embiids readily fold transversely and slide for- 
ward toward the head (Fig. 23), thereby reducing 
likelihood of a snag, or "barb-effect." 

Although the wings usually fold upward and 
cephalad across their midline, they can bend at al- 
most any point and may even irregularly crumple. 
Such flexibility appears to have been accomplished 
through desclerotization of most of the longitudinal 
veins behind the radial blood sinus (RBS), notably 
the media (M) and cubitus (Cu). Perhaps reduction 
of plication is also involved, as suggested by the fact 
that vein cuticularization is almost entirely confined 
to the dorsal membrane with only blood sinus veins 
evident on both wing surfaces. 





B 

Figure 22. A. Hind tarsus of first instar and adult of male 
OUgotoiua nigra. B. First instar and adult hind tarsus of 
female Haploembia soUeri (both showing medial papilla, 
lacking in O. nigra). 



26 




OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 

11^ 



m 



'ms^' 



"*L^ 




* 



Figure 23. Wing flip of a reverse-moving male. New genus and species, Oligotomidae. Thiailand. 



The great flexibility of such wings leads one to 
question how they could serve as flight organs. The 
answer seems to lie in the functioning of four blood 
sinus veins, the most important of which follows the 
course of the anterior radius (RBS - RA). Second- 
ary blood sinuses include the subcostal (ScBS - Sc), 
the cubital (CuBS), and the anal (ABS = A) (Fig. 
24). These sinuses con'espond in position to the as- 
sociated veins except for the terminal halves of CuB S 
and ABS which do not contain tracheae. Through- 
out the order the sinus veins are cuticularized on both 
the dorsal and ventral wing membranes. 

In effect, each sinus vein is a dark, glossy, dor- 
soventrally-tlattened sac that is tapered and perhaps 
closed distally. When wings are in repose, sinus sur- 
faces are flat but during "excitement" preceding 
flight, hemocoelic pressure must increase and. as the 
wings extend laterally preparatory to flight, 
hemolymph ("blood") flows into the sinuses causing 
their surfaces to become convex. Microscopic ex- 
amination of a turgid sinus shows that hemocytes 
synchronously pulsate with the beat of the dorsal 
blood vessel and do not perceptibly move distad. 

Together, the turgid subcostal, cubital and anal 
sinuses appear to function as tines of a fork that stiff- 
en the wing's base, while the radial blood sinus (RBS) 



stiffens almost the entire length of the wing's lead 
edge. When an embiid alights, and the wings return 
to repose over the body, it is probable that a crimp- 
ing of the blood sinuses occurs in the axillary region. 
Hemolymph then gradually oozes back into the body 
cavity. 

One could regard, especially the longitudinal wing 
veins of all insects, as narrow sinuses into which co- 
elomic blood pressure extends as a means of expand- 
ing wings from pad to adult form. Following a teneral 
period, during which these veins gradually cuticu- 
larize, blood pressure into them must decrease, or 
completely cease except for pressure in the axillary 
area which remains to extend wings for flight ("take- 
off'), or defensive, or sexual, display, among insects 
which fold their wings over the body when they are 
not in use. 

In contrast, blood sinus veins of embiids enable 
blood pressure to continue on and off throughout the 
adult life of all species with alate males. Over time, 
such veins broadened and sclerotized to become dis- 
tally-closed sinuses characterizing wings of all spe- 
cies of the order Embiids apparently are the only 
insects — indeed the only animals on Earth — capa- 
ble of temporarily stiffening, otherwise flexible. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



27 



RP MA C, 



Sc or ScBS 




AorABS 



FOREWING 



CuP CuA; 



CuA, 




--••"5^-A 



HINDWING 

Figure 24. Wing venation of a plesiomorphic embiid, Clothoda longicauda Ross (Clothodidae). ScBS, RBS, CuBS 
and ABS are symbols for blood sinus veins. RML = radius marginal line. Note; MP vein is anomalous in hindwing. 



The fact that the sinus veins are similarly-deve- 
loped throughout the order indicates a great antiqui- 
ty for the specialization, but one preceded by 
evolution of the silk-spinning ability and increased 
survival associated with the complete confinement 
in silk galleries. Such unusual wing adaptations must 
have been initially perfected in a single species for it 
seems unlikely that they could have evolved with 
identical complexity more than once. Furthermore, 
primary selection pressure for wing flexibility was 
most likely on adult females, not adult males, to adapt 
them more perfectly to gallery life. The adult life of 
a male is too short and its prime biological function — 
simply a non-feeding "vehicle" for delivery of sperm 
and genetic diversity — is performed so quickly that 
there would not seem to be sufficient selective ad- 
vantage for males to have been the primary target for 
this remarkable wing specialization. Adult females, 
however, must live long enough to mature eggs, ovi- 
posit and guard eggs and early-instar nymphs. Thus, 
a species more significantly benefits from adapta- 
tions which foster quick, predator-avoiding move- 



ment of females in narrow silk galleries. The evolu- 
tion of alternating wing flexibility and stiffening rep- 
resents one of the order's first "attempts" to ease 
backward movement. 

Venation 

Strength (cuticularization) and completeness of 
venation (plesiomorphic features), are greatest in spe- 
cies with largest body size (also plesiomorphic). 
Complex venation apparently is related to the obvi- 
ous need for more extensive blood and air distribu- 
tion in larger wings, especially in the pads of large 
nymphs during development. 

Wings of Clothoda longicauda Ross (Fig. 24), one 
of the order's most plesiomorphic species, may be 
used for interpreting venation. Most past work on 
Embiidina, including mine, used Comstock-Needham 
nomenclature of veins, as summarized by Comstock 
( 1918). Based on tracheation, a modified nomencla- 
ture is here adopted. I am aware that many entomol- 
ogists since Comstock have rejected tracheation as a 
basis for interpretation; however, in embiids there is 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



consistent correlation between tracheation and ve- 
nation, both in pads and in fully developed wings. It 
should be noted that tracheae aren't easily seen, if at 
all, in wings of dead, dry specimens but are most 
apparent in freshly killed individuals, especially those 
still teneral (Fig. 25 A, B). 

The costa (C) forms the anterior wing margin and 
terminates at the apex of the anterior radius (RA). 
The subcosta (Sc) is short, cuticularized, terminates 
within the wing's basal fourth, and probably func- 
tions as a blood sinus, hence the symbol ScBS. The 
anterior radius (RA), the wing's most prominent vein 
and, most significant blood sinus (RES), stiffens the 
wing's lead edge. It originates strongly near the wing 
base and parallels the costal margin almost to the 
wing's apex, at which point it tapers and usually 
curves downward to join RR Especially in apomor- 
phic species of Anisembiidae, RA slants toward the 
costal margin well before the wing's distal curvature. 

Throughout most of its length RES is exception- 
ally broad, sclerotized, glossy, darkly melanized and 
bordered, except at basal fourth, by peculiar streaks 
(radius margin lines, RML), having a fleshy, granu- 
lar appearance, and usually are brick red in color. 
Images taken by SEM show the surface of these lines 
to be irregularly wrinkled, perhaps to accommodate 
alternate turgescence and flattening of RES. The ra- 
dial blood sinus and its margins would be an inter- 
esting subject for detailed study. 

Near the base of RES a short stem juts caudad 
and then immediately extends distad (RP -i- MA) for 
almost a third of the wing's length, at which point it 
appears to fork. However, the anterior branch of the 
"fork" is simply a continuation of RP, which never is 
forked in any species of the order. Thus, the basal 
portion of this "vein" really represents a fusion of RP 
and MA. Superficially, especially in long dead spec- 
imens, this composite stem appears to be a single, 
cuticularized vein; however, two parallel tracheae 
within it have separate wing-base origins, a condi- 
tion which prevails in wings and wing pads of all spe- 
cies of the order (Fig. 26). The posterior branch of 
the fork, once designated R4+5, is here regarded as a 
continuation of the anterior branch of the media (MA). 
This vein usually is forked, forming MAi+: and 
MA3+4, especially in plesiomorphic taxa of the order. 

The stem of the "media" is fused to the anterior 
edge of the sclerotic base of the cubital blood sinus 
(CuES). Separate tracheae, within this stem angle 
abruptly forward toward the stem of RP and then sep- 



arate to form MA and MP. MP then extends to the 
mid-margin of the wing and is very rarely forked. 

The trachea of the cubitus (Cu) at first parallels 
those of MA and MP within the extreme base of the 
cubital blood sinus (CuES). but then angles caudad 
within the sinus. At the sinus' mid-length the tra- 
chea and its vein emerge from the sinus and parallels 
MP to the wing margin. In plesiomorphic embiids, 
or those with exceptionally large wings, the cubitus 
may be multibranched. I tentatively regard the basal 
branch of Cu as CuP. 

Eeyond this diversion of Cu. the cubital blood 
sinus continues its broad, tapered, diagonal course 
to the wing's hind margin. No trachea follows the 
distal half of the sinus. The hyaline stripe between 
CuES and the anal blood sinus (AES) tends to crease 
and this suggests that it may be equivalent to the cla- 
val suture which delimits the anal fold in wings of 
certain other insect orders. It tends to fold upward 
and forward in embiids. 

As stated before, the anal (or vannal) area of 
embiid wings is much reduced, but it has a dark cen- 
tral line, an anal blood sinus (AES), within which 
one can see the unbranched trachea of the anal vein 
(A). Kukalova-Peck (pers. com.) prefers to desig- 
nate A as AA because, in many other insects, there is 
a posterior branch of A, therefore an AP vein. There 
is always a cross- vein between A and the base of the 
cubital blood sinus. Kukalova-Peck regards this as 
an "anal brace." 

The hindwing is similar to the forewing but al- 
ways is shorter, broader and certain veins, such as 
MP and Cu, may be less strongly represented. In 
some species of Archembia Ross, the anal area is 
slightly more expanded than in the forewing (Fig. 27), 

Cross-veins may be highly variable in position 
and number within a species and may even differ in 
the left and right wings of a single individual. How- 
ever, their general positions and number is rather con- 
stant within a species, or even a genus. In Oligotoma, 
for example, cross-veins seldom if ever are present 
behind RP. There seems to be no regularity of cross- 
\ein position which would justify nomenclature for 
cells they delimit. 

The upper and lower wing surfaces are densely 
clothed with small, short hairs commonly called 
microtrichiae which, having no apparent basal sock- 
ets, appear to arise as direct outgrowths of the wing 
membrane. The entire outer margins of the wing and 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



29 




HINDWING 




B 



FlGI-iRE 25. A. Dark field illumination of wings of freshly killed "Embki" surcoufi Navas (Embiidae) showing tracheae 
(not visible in "dead" wings). B. Tracheation of teneral Avchembia batesi (McLachlan) (Embiidae). Photo also shows 
enlarged anal area of the wing occurring in some species of Archeiubia Ross. 



30 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 




Figure 26. Tracheation of forewing pad of penultimate 
instar "Embia" siircoufi Navas. 



courses of all veins, even those unsclerotized, bear 
rather large setae (macrotrichiae) arising from prom- 
inent sockets. They are particularly large along the 
costal margin. Large setae are also present on the 
ventral wing membrane but are fewer in number and 
without definite arrangement. 

The wings of Pavarhagadochir Davis (Fig. 28) 
of the Embiidae exemplify the more apomorphic (= 
reduced) wing venation found in most Embiidina. 
In such venation the apex of MAi+: and MA3+4 and 
all of MP are unsclerotized. each vein traceable only 
by its row of macrotrichiae and pigment stripe. 

The wings of Enveja beqiiaerti Navas (Fig. 29) 
represent contrasting, perhaps plesiomorphic. vena- 
tion in which all veins are heavily sclerotized, but 
the cubitus isn't forked. 

In some species of the apomorphic family Ter- 
atembiidae. as well as in other taxa. small body size 
correlates with vein reduction, including all veins ex- 
cept those functioning as blood sinuses (Fig. 31). 

In a common type of reduction. MA always is 
simple (Fig. 30). All species of unrelated families 
Anisembiidae and Oligotomidae have such reduc- 
tion, a reduction which sporadically occurs in sev- 
eral other distinct evolutionary lines, such as within 
Embiidae, Notoligotomidae and Teratembiidae. 

The most apomorphic wing of the order is found 
in a South African new species of Teratembiidae. Its 
wings are very small, slender, with all veins except 
the blood sinuses obsolete, and the wing margins have 
especially long setae. Such thysanoptery parallels 
the tendency of the smallest species of various insect 
orders (e.g., certain parasitic wasps, small Tri- 
choptera, some microlepidoptera and ptiliid beetles) 
to have slender, fringed wings. 

Although there is much convergence in wing ve- 
nation in embiids. \ enational characters have impor- 



tant, supplemental value in the definition of species, 
genera, and even families. It is doubtful, however, if 
wing characters can ever be used as the primary ba- 
sis of phylogenetic conclusions. 

Perhaps, because of the unimportance of flight 
in evasion of predators, embiid wings exhibit con- 
siderable random, often anomalous, intraspecific vari- 
ation. The most extreme, yet consistent, wing 
anomaly occurs in an Amazonian new species of 
OUgembia Davis, which has normal forewings but, 
not even a trace of hindwings (Fig. 32), and the met- 
athorax is reduced to the size of an abdominal seg- 
ment. However, very closely related species from 
the same region have normal hindwings and thus the 
hindwing atrophy is of no significance in systemat- 
ics. If. however, a comparable character appeared in 
certain other insect orders, it might become the basis 
for proposing a distinct higher taxon. 

Embiid wing anomalies seem to illustrate a law 
in biology which may be expressed, as follows: if an 
anatomical feature is not vital to survival or repro- 
duction, it may be subject to much anomalous varia- 
tion within a taxon. Thus, in flight-dependent insects, 
such as most Diptera and Hymenoptera, wing fea- 
tures are relatively constant and are, therefore, de- 
pendable characters in systematics. The converse 
appears to be the case in embiids because their flight 
has little or no adaptive value. 

Wing pigmentation 

Apparently pigmentation of embiid wings always 
is confined to the upper membrane, the ventral being 
completely hyaline except for dark "imprints'" of the 
blood sinuses. Alternating longitudinal dark stripes 
and hyaline intervals, although faint in some species, 
are characteristic features of the upper membrane of 
all embiid wings except possibly those of Burmitem- 
bia venosa Cockerell, an Eocene (?) amber fossil from 
Burma. 

The veins and/or their macrotrichiae are centered 
in the dark stripes; the intensity, width and marginal 
definition of which are consistent within a species. 
In turn.-such melanization correlates with the overall 
pigmentation of the male. In arid regions many spe- 
cies disperse noctumally and generally are pale tan 
with wings correspondingly pale with faint venal 
stripes and broad hyaline intervals, the margins of 
which are often indefinite and irregular. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART I 

RR.S 



31 




MA, 



MA, 



Figure 27. Wings of Archembia n. sp. (Embiidaej showing narrow hyaline stripes and broad anal area (a plesiomorphic 
condition). 



RBS 




RP 



MA, 



CuBS 



MA 3+4 



Cu„ 



MP 



Figure 28. Forewing of Pararhagadochir trinitatis (Embiidae) showing reduced anal area. 




ABS 



MA3^4 



Cu„ 



Figure 29. Forewing of Enveja bequarti showing strong venation and white cross-veins. The wing's costal and hind 
margins are golden. 



32 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



RES 



ABS 




MA 



CuBS 



Cu,, 



Figure 30. Reduced vein-strength of the forewing and venation of a species of Clielicerca (Anisembiidae). The unforked 
MA vein characterizes many species and entire families, e.g., Anisembiidae and Oligotomidae. 



RBS 




Figure 3 1 . Reduced vein-strength of the forewing of Teratembia geniculata (Teratembiidae). Many African species of 
this family have vein MA unforked. 



' t. ^ 




Figure 32. Complete atrophy of metathora.x and hindwing loss in a new Brazilian species of Oligembia (Teratembiidae). 
Terminalia characters indicate, however, that this isn't a very distinct species. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



33 



There are exceptions to this, however, for melan- 
ism of some species occurring in certain seasonally- 
arid regions (e.g., Embia Latreille spp., in the 
Mediterranean region) is often associated with nup- 
tial dispersal during cool weather following winter 
and early spring rainfall. In such cases melanism 
may foster rapid body warming in the sun, enabling 
more rapid dispersal movement, thereby reducing 
predation hazard. 

Especially in the humid tropics, wings of darkly- 
pigmented males (e.g., Ptilocerembia Friederichs 
spp.) often have a beautiful, metallic blue or laven- 
der luster. This is especially intense on the veins and 
diminishes toward the hyaline stripes. 

In some species wing pigmentation may contrib- 
ute to a mimetic appearance. Blackness causes the 
wings to resemble elytra of aposematic beetles, such 
as Pyrochroidae, which, like their embiid mimics, 
usually have a reddish prothorax and black elytra. 
In another type of mimicry, extensive golden mar- 
gins of the wings (e.g., Enveja Navas spp.) result in 
a resemblance to chemically-repugnant lycid beetles 
occurring in the same habitats. 

Curiously, cross-veins often are conspicuously 
white when crossing hyaline intervals and darkly pig- 
mented while crossing the dark stripes. The result- 
ant white, cross-slashing of embiid wings is 
characteristic of richly-pigmented, diurnal males 
occurring in humid environments. 

As mentioned before, brick red or pink, "granu- 
lar" lines (RML - radius marginal lines) bordering 
the radius blood sinus characterize all embiid wings. 
This subcutaneous granular pigmentation may also 
be present in the costal margin and in veins RP and 
MA, as in Pararhagadochir Davis (Fig. 28). Each 
side of a radius margin line may be pale. Sometimes 
the granular red lines extend into adjacent longitudi- 
nal veins and cross-veins. 

Reflecting on the signifance of the universal strip- 
ing of embiid wings, I have concluded that wings of 
ancestral species must have been uniformly dark and 
that the hyaline intervals evolved in lines of weak- 
ness which fostered longitudinal plication. Observed 
obliquely, embiid wings display slight ridges corre- 
sponding to the veins, and furrows correlated with 
the hyaline intervals. Unlike most insect wings, those 
of embiids do not have a pattern of positive and neg- 
ative longitudinal veins; the ventral membrane lacks 
cuticularized veins. However, at least in some gen- 



era, e.g., Clothoda Enderlein, veins are pale in color 
on the ventral membrane. 

In some clothodids, such as Clothoda nobllis 
Enderlein and Antipaluria marginata Ross, the wing's 
costal margin is white. In some species of Anisem- 
biidae and Teratembiidae extreme apices of the wings 
are white. 

Wing expansion following ecdysis 

During most of the penultimate instar, wing pads 
of males are flat with venation identical to that of 
adults (not zig-zagged, for example, as in some Ple- 
coptera). Nearing ecdysis, the pads become thick- 
ened, convex and opaque white. When the exuviae 
is shed, the pads at first retain the fleshy shape and 
jut out from the thorax at about a 30° angle. This 
probably assists flow of blood into the pads. 

Within ten minutes the projected pads begin to 
flatten and expand from the costal to anal margins. 
Then they gradually assume the normal, repose po- 
sition over the dorsum of the thorax. Periodically, 
the embiid wriggles and rotates its body. Concur- 
rently, the abdominal terminalia are distended and 
the cerci project laterad at 45° — perhaps due to an 
increase in hemocoelic pressure throughout the body. 

In about twenty minutes the basal half of each 
wing has fully expanded and flattened, the distal half 
remaining as naiTow and as fleshy as at the time of 
ecdysis. In about thirty additional minutes the entire 
costal margin has expanded and only the distal ex- 
tremity of the posterior margin remains fleshy. This 
condition prevails for another thirty minutes after 
which the wings are fully expanded but remain white 
with veins paler than the intervening membranes. 

About two hours after ecdysis the wings have at- 
tained their definitive shape and thickness and their 
pigment stripes and hyaline intervals are faintly ap- 
parent. Seven hours after ecdysis the wings are gray 
in tone and the body and leg pigmentation is well 
developed. After about twenty-four hours wing pig- 
mentation, or cuticularization, is completed and the 
male has ingested his exuviae. The male tends to 
remain in one position for at least another day fol- 
lowing ecdysis. 

Wing articulation 

I am indebted to Jannila Kukalova-Peck who, dur- 
ing a visit to my laboratory (1999), greatly improved 
my treatment of wing articulation. Details, presented 



34 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



in Figure 33A. will be elaborated and possibly cor- 
rected, in one of her future publications. 

Articulation of the hindwings is similar to that of 
forewings. There is. however, a trend toward a slight- 
ly weaker representation of sclerites which is un- 
doubtedly correlated with the wing's smaller size. 

Flexion of wings 

In repose the wings lie flat over the back (Fig. B) 
much as in termites, zorapterans and stoneflies. The 
anal area of the wing is nearly obsolete and is repre- 
sented only by a small, basal, posterior comer which, 
in a fully-flexed wing, folds inward beneath the wing 
surface. The important fold occurs at the wing base. 
I have observed the mechanism of this basal flexion 
in both living and dead specimens of Oligotoma ni- 
gra and noted movements of the sclerites. as follows: 
the fulcrum, or pivot, of the flexion is anterior; being 
at the point of aticulation of the anterior part of the 
first axillary sclerite with the base of the subcosta. 
An imaginary line drawn from this point through the 
basal articulation of the third axillary, and still an- 
other from it through the point of contact of the apex 
of the third axillary with the posterior angle of the 
anal band, delimit a narrow triangular area. As the 
wing returns from the completely extended position 
to rest over the back, folds occur along these lines 
and the area becomes completely inverted. During 
this movement (or the reverse) only the three axil- 
lary sclerites change position while the other parts of 
the wing-base remain stationary. During flexion the 
first axillary rotates against the anterior wing pro- 
cess through a 90° arc. The second axillary becomes 
upright and the third is entirely inverted. The anteri- 
or membrane is stretched around the fulcra] point and. 
finally, at least half of it comes to lie parallel with 
the side of the body. 

The flexion of the wing thus seems to correspond 
to that of many other insects. One point requiring 
further investigation, in the light of the order's wing 
peculiarities, is the possible control of blood circu- 
lation in the large radial blood-sinus (RBS). and oth- 
er blood-sinus veins by means of movements in the 
wing base. This may occur, at least in the case of 
RBS. b\' simple pressure of the anterior membrane 
against the fulcral point. There is a small strength- 
ened point in the membrane opposite this fulcra! point 
which ma},- fit across the place of strongest contact. 



Flight 

Because predator-avoidance especially depends 
on remaining within silk galleries, flight did not 
evolve as an important means of defense or dispers- 
al. Adult males are slow to take flight and do not 
readily fly away from a disturbance. They are more 
likely to run away. In preparation for flight, a male 
rises high on his forelegs, at times lifting them off 
of the substrate, the head may bob up and down, and 
the antennae may vibrate and rapidly twirl. Flight 
distance usually is short, perhaps seldom exceeding 
a meter and, soon after alighting, there may be a rep- 
etition of the pre-flight and flight behavior. At times 
take-off follows a short run or a hop. 

Flight is a swirling, aimless, fluttering with ap- 
parently no more directional control than that of nup- 
tials of most species of termites. However, males of 
nocturnally-dispersing species fly toward artificial 
lights. It may also be assumed that they can direct 
their flight to a gallery containing a receptive fe- 
male. Especially in flight, diumally-dispersing males 
of certain species can easily be mistaken for various 
aposematic beetles, such as some lycids, and 
pyrochroids. which also serve as models for mim- 
icry of many other insects. 

Reduction and elimination of 
wings 

Obviously, the ultimate adaptation for rapid re- 
verse movement in galleries is complete wing elim- 
ination, now universal in females, through neoteny. 
This also has occurred independently in males on 
almost every evolutionary line within the order. A 
similar neotenic "solution" developed among work- 
er and soldier termites. Alate termites, however, break 
off no-longer-needed wings prior to copulation and 
a return to gallery life. 

Due to friction, it is probable that more winged 
embiid males are caught by predators at the extremi- 
ties of galleries than are apterous, or subapterous, 
members of a colony. Thus, mutations resulting in 
wing-loss, or reduction in their size, are likely to be 
selected. Another, and perhaps more significant fac- 
tor, is that alate individuals are more likely to make 
hazardous flights out of protective galleries and 
thereby become exposed to predation and other haz- 
ards, such as desiccation. The fact that aptery is not 
yet universal in males may be explained, as follows: 



ROSS; EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



35 



tegula 



basivenalii 
BR.BM.B 



' C 



3rd axillary sclerlte 

posterior wing 
process 



RBS 



RP -MA 




CuA 



Figure 33. A. Wing attachment of forewing of a plesiomorphic embiid, Archembia kotzbaueri (Navas), Embiidae. 
Nomenclature based on studies of Dr. Kukalova-Peck (pers. com.). Explanation of symbols: Sc - subcosta; BSc.BR. 
BM. BCu. BA = subcostal, radial, medial, cubital, and anal basivenales; medial plate including medial (FM) and cubital 
(FCu) fulcalares, is almost completely reduced. FA = minute anal arm of third axillary sclerite. The symbol AA (anterior 
anal vein) is used because there is a tiny anal vein just caudad of the base of AA (not expressed in this species). AABS = 
anterior anal blood sinus; CuBS = cubital blood sinus; RBS = radial blood sinus. Note that in this drawing the axiilaries 
are somewhat spread apart and that the distance between CuBS and AABS is exaggerated. 



1st axillary sclerite 
2nd axillary sclerite 




anterior 

wing process 



posteriar 

wing process 



axillary 
cord 




B 



Figure 33. B. Dorsal aspect of wing attachment of Oligotoma nigra. C. Pleural aspect of wing attachment. 



36 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



1. Selective pressures favoring wing-loss muta- 
tions are too short in duration to be effective. Life of 
adult males is undoubtedly short (alate males take 
no food) and they probably mate before the disad- 
vantage of wings is fully felt. In other words, imper- 
fect adaptation of alate males to gallery life is of no 
importance to a species once males have mated. 

2. It is also possible that factors which might 
result in brachypterism, or apterism, of males have 
not yet appeared for test in all species. 

3. Wing possession may be less disadvantageous 
in favorable environments, such as tropical 
rainforests. In such habitats apterism of males is very 
rare, whereas it is very common in regions with a 
long dry season. It is thus likely that alate males dis- 
persing in arid regions not only face exposure to ad- 
verse cUmatic conditions, but also are more visible 
to predators seeking prey in exposed arid habitats. 

4. Wide dispersal of advantageous genes is se- 
lected through retention of wings. 

It may be that universal wing-loss in females was 
not entirely due to selection against wing-possession 
per se, but instead for a need to also eliminate pro- 
jecting ovipositor structures which might have had 
an even greater barb-effect than wings in reverse 
movement. One could say that females lost their 
wings as part of a nymphalization "package" that 
resulted in a loss of most protruding adult append- 
ages by cessation of development at an early nymphal 
instar, perhaps not later than the second, before even 
buds of such appendages make their appearance. This 
neotenization was probably effected by mutations 
influencing the secretion of the juvenile hormone. 

Such a major specialization of wing structure and 
function must have occurred long after the order 
evolved most of its other peculiar specializations be- 
fore fragmentation of Pangaea. It would be difficult 
to explain the present widespread distribution of the 
order if females had been apterous and needed to 
hazardously extend their range afoot outside of pro- 
tective galleries. Both sexes must have been alate 
during the major evolutionary and distributional his- 
tory of the order and female apterism must have later 
convergently occurred on all evolutionary lines. The 
appearance of more aggressive predators, such as 
ants, could have been a factor favoring apterism and 
confinement to galleries. 

Apterism in males (Figs. 34, 35, 36) also has 
independently appeared many times within the order. 



but in varying degrees. Females can deposit eggs in 
galleries without specialized oviposition structures; 
males, however, with the exception of those of a new 
family from Afghanistan, must have well-developed 
genitalia and mandibles (often used as head claspers) 
to insure copulation. Thus, even in highly neotenic 
males, such as those of the Australian family 
Australembiidae, neoteny mostly retards wing 
development while, apparently through "tissue 
competence," the genitalia and head (to a lesser 
extent), become fully adult. In some species of 
Australembiidae, however, the head of adult males 
is variable, an occasional individual within a 
population has a head indistinguishable from that of 
a nymph, or of an adult female, and continues to feed. 

Males of many species gain advantages of 
apterism through degrees of thoracic neoteny. As a 
result, males of some species have only wing buds, 
or pads, in various states between the extremes of 




Figure 34. Adult neotenic, apterous male (left) and fe- 
male of anew species of A'eor/iaga<ioc/!!r Ross (Embiidae) 
from an arid region of Nicaragua. Unlike its blackish 
congeners, this species is pale ferrugineous due to its sub- 
terranean habits. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



37 




Figure 35. Neotenic, completely apterous adult male of 
Haploembia solieri (Oligotomidae) (slide preparation, body 
length 10.0 mm). Endemic to the Mediterranean region. 
Note; tong-like mandibles for grasping head of females 
during mating. 




Figure 36. Neotenic apterous adult male of Electwembia 
antiqua (Pictet) (Embiidae). Baltic Amber, Hamburg Geo- 
logical Museum, body length about 10.0 mm. This fossil 
demonstrates antiquity of neoteny in males and close re- 
semblance to modem species. 



full wing pad development and no trace of pads what- 
soever. All this is probably due to different levels of 
secretion and timing of juvenile hormone. In some 
species there may be percentages of apterous, 
subapterous, micropterous and alate males within a 
species' population, or in those of certain geographic 
populations of a single species. 

In arid environments apterous males are more 
likely to remain within the parent colony, mate with 
a sister, and thereby inbreed the wingless trend or 
condition. Because of their greater ease of move- 
ment, apterous males should survive in greater num- 
bers and eventually male apterism could become 
universal within a given population. Conversely, in 
damper, more benign environments, any trend toward 
male apterism might be swamped out by random 
matings of alate individuals which are more likely to 
survive flights from colony to colony. 



Some families include genera and species which 
have radiated into marginal environments and thus 
have apterous, or subapterous males. Therefore, 
apterism in males must be used with great caution as 
a character in systematic studies for it is most often 
environmentally related. It is interesting to note that 
one of the oldest known fossil .species, Electwembia 
antiqua (Pictet) of Baltic Amber (Ross, 1956). is com- 
pletely apterous (Fig. 36) and this suggests that the 
ancient land surface which supported the "amber 
forest" might have experienced a long dry season. It 
is possible that it was once in what is now a Mediter- 
ranean, seasonally-dry global position, a teixain which 
has since drifted northward into a colder, wetter lati- 
tude. 

Ultimate in the trend toward almost complete 
neotenization of males is in a peculiar, undescribed, 
subterranean species occurring in the desert steppes 
of western Afghanistan. In this species males not 



38 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



only have lost all traces of wings but also have com- 
pletely nymphoid bodies and abdominal terminalia 
(except for tiny rudiments). 

There is an interesting correlation between wing 
size and body proportions. Relatively large wings 
are characteristic of light-bodied, slender, small- 
headed males of species occurring at higher eleva- 
tions of damp, equatorial mountains. In contrast, 
shorter, narrower wings are possessed by robust, 
larger-headed males living at lower altitudes and in 
semi-arid regions. Except for slight venational diver- 
sity, often involving vein-desclerotization. the wings 
of embiids are remarkably similar in all species. 

Abdomen 

The abdomen is slender, usually as long as head 
and thorax combined. In nymphs, adult females, and 
apterous adult males of some species, it is circular in 
cross-section. However, in alate adult males of most 
species it is dorso-ventrally flattened due to reduced 
content, all food having been excreted during the 
penultimate nymphal stage. Fat storage is limited, 
and internal reproductive organs are much smaller 
than those of females. 

Ten abdominal somites are conspicuous in both 
sexesbut vestiges of the 11th and 12th persist. Basic 
somatization is most apparent in nymphs and adult 
females whereas that of adult males is confused by 
complexity of external genitalia, especially in 
apomorphic taxa, as illustrated (Figs. 44—53). 

The first abdominal tergum of nymphs and adult 
females, a simple plate without an acrotergite. closely 
contacts the metathoracic scutum. In alate males an 
extensive, medially-cleft acrotergite is fused to the 
metathoracic scutum. Terga of somites two through 
eight are similar to each other but the ninth is much 
shorter, broader, extends ventrad down each side of 
the abdomen and almost contacts the outer margins 
of the ninth sternum, or hypandrium (H). In nymphs 
and females the tenth tergum is large, triangular, and 
its outer basal angles extend ventrad to the sides of 
the ninth sternum (Figs. 37, 38). 

Matsuda (1976) regarded the produced apex of 
the tenth tergum as the supra-anal lobe fused to the 
tenth tergum. I have concluded that only the small, 
weakly sclerotized area just beneath the apex of the 
tenth tergum is a vestige of the eleventh (labelled 
epiproct in Figs. 37 and 38). In females of some 
species this vestige is separated from the apex of the 



tenth by a transversely wrinkled, non-setose, inter- 
somital membrane and the vestige bears its own se- 
tae. Edward L. Smith informs me that intertergal 
muscles are attached to this sclerite. It, and the apex 
of the tenth tergum. develop into significant terminaha 
structures, e.g., the epiproct (EP) and the medial flap 
(MF), prominent in adult males of many species. 

Located just beneath the lateral margins of the 
first eight abdominal terga are elongate laterotergites, 
each of which has a spiracle in the anterior end. In 
many species the laterotergites of these somites are 
divided into two sections, the posterior of which usu- 
ally is much smaller (Fig. 38). Spiracles and 
laterotergites are absent on somites nine and ten, the 
positions of the latter being filled by latero-ventro 
extensions of the terga. 

In many apomorphic embiids, such as species of 
Oligotoma, the first sternum is small and triangular 
but in plesiomorphic genera, such as Clothoda and 
Embia, it is larger and more transverse. Stemites of 
somites two through seven are nearly equal in size 
and form, each being subquadrate with a narrower 
base. In adult females stemites of somites eight and 
nine, which are associated with the vulva, are vari- 
ously modified and will be separately discussed. 

On either side of stemites three through eight there 
are narrow, elongate sclerites similar in shape to the 
laterotergites. Apparently these are abdominal 
pleurites. They are almost entirely absent on the first 
two somites, being represented only by two setae 
adjacent to the stemite of the second somite. Pleurites 
are absent on somites nine and ten. Inevitably, there 
are difficulties in interpreting terminal abdominal 
somites. Using Snodgrass (1935) as an authority, I 
have decided that the paraprocts are structures of 
somite nine. Edward L. Smith (pers. com.) believes 
that the paraprocts are hemistemites of somite ten. 
Matsuda (1976), however, regarded them as stmc- 
tures of somite twelve, suggesting that somite eleven 
is greatly reduced with cerci remaining as its only 
elements and. therefore, that structures of somite 
twelve immediately follow those of somite ten. I 
am not prepared to question these conclusions and 
will simply endeavor to correctly homologize 
terminalia characters without being overly concemed 
with their somital associations. 

The cerci of nymphs and adult females are simi- 
lar throughout the order and seemingly comprise only 
two segments (I realize that these are properly termed 
flagellomeres but, for simplicity, I use the term seg- 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



39 



ment). Cercus-basipodites, or coxites, may be re- 
garded as basal segments of three-segmented cerci. 
Each basipodite forms an almost complete fleshy ring 
around the base of a cercus. The cercus segments 
usually are elongate, cylindrical and often unevenly 
sclerotized. Except for a relatively large nerve, most 
of the content of a segment appeal's to be fat. The 
cercus muscles are attached to the basipodites which, 
in the left cerci of adult males of many species, be- 
come fused to the base of the basal segment. In 
plesiomorphic species the derm of the cerci is evenly 
sclerotized. but in some apomorphic taxa it may be 
almost entirely membranous. As a species character 
the distal segment may be contrastingly pale or white 
due to the color of tissue within a transparent derm. 

The cerci bear setae of two types. Most numer- 
ous are ordinary, tapered setae of various sizes which 
arise from simple, circular sockets. These occur on 
all surfaces but may be especially dense on the inner 
faces of the basal segments. In males such density 
may augment copulatory grip. The second type of 
seta is finer, less tapered, arises from a rosette-type 
of socket, or pit, and is most common on the usually- 
less-sclerotized outer side of the basal segments. In- 
variably throughout the order, the distal segment has 
only one such seta on its inner side in a species-char- 
acteristic position. Such setae are present in many 
other arthropods and are often called trichobothria. 
Probably both types of setae are mechano-receptors 
providing tactile guidance, especially during reverse 
movement in the galleries. 

External genitalia of females 

Because of neoteny, the external genitalia of fe- 
males are underdeveloped. However, some species 
possess buds of gonopophyses which might attain 
adult form if females completed development, as they 
did during their pre-neotenic evolutionary period. 

Reduction of genitalia is possible because of the 
simplicity of oviposition. A female merely attaches 
eggs to a surface within the galleries, or on a silk 
substrate, therefore no special structures are required 
to insert them into a substrate. Furthermore, vital 
reverse movements to escape predators would be 
slowed, or arrested, if ovipositing structures protruded 
and snagged against silk gallery walls. 

The terminal abdominal terga, paraprocts, and 
cerci of adult females (Figs. 37, 38) are identical to 
those of nymphal instars. The only external evidences 



of maturity, besides the open vulva, are slight modi- 
fications of the eighth and ninth sternites. In 
OUgotoma (Fig. 33), and many other genera, there is 
no trace of valvulae, but between the eighth and ninth 
sternites there is a slight, transverse, translucent ridge, 
or carina. This elevation is subject to much modifi- 
cation within the order; for example, in many spe- 
cies of Embia its caudal side has two deep fossae 
with glossy, sclerotic surfaces. The ninth stemite 
usually has a baso-medial notch which varies from a 
membranous condition to the dark, glossy, concave 
sclerite found in some species of Embia. 

In other genera, notably MetoUgotoma Davis, ru- 
diments of valvulae are quite conspicuous. In Meto- 
Ugotoma (Fig. 38) the eighth sternite is small and 
lies beneath two blunt, fleshy lobes, or pads, which 
bear small, rudimentary sclerites. These appear to 
be rudiments of the first valvulae. Overlapping the 
anterior margin of the ninth sternite, a prominent, bi- 
lobed, non-setose ridge (probably a homolog of the 
membranous ridge described for OUgotoma) is 
present which may be a specialized rudiment of the 
second valvulae. The lobes themselves may be rudi- 
ments of the second valvulae and the low connecting 
ridge represents anterior intervalvula. The ninth ster- 
nite is small and has a lai'ge, quadrate membranous 
area in the basal half. The pouch-like development 
of this area, described for OUgotoma. is well devel- 
oped in MetoUgotoma and is partially concealed by 
a vestige of the base of the second valvulae. The 
aperture of the accessory gland may be located in 
this pouch (Snodgrass, 1935, fig. 314B). 

Although species of Clothoda have the most ple- 
siomoiphic males of the order, adult females of the 
genus lack even traces of valvulae, the ninth ster- 
num being simple and lacking a basal pouch. 
Throughout the order interspecific variation occurs 
in sclerotization, pigmentation, and vestiture of fe- 
male genitalia and is of potential value in systematic 
studies, at least at the species level. 

Internal genitalia of males 

Probably the distal portion of the ejaculatory duct 
is projected into the vulva during copulation but it 
rarely, if ever, has sclerotic rigidity — an aedeagus. 
Exceptions to this ai'e especially apparent in the ge- 
nus Enveja Navas (Fig. 50), and to a lesser degree in 
most genera of Anisembiidae, and in some species 
of Oligotomidae: but this is merely limited to sclero- 
tization of the duct walls. In mv figures, labelled 



40 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 

7th torgum 8lh tergum 9th lorgum loth lergum 



- cercus- basipodite 




laterotergites 1st valvula 



accessory gland aperture 



Figure 37. Abdominal terminalia of adult female Oligotoma nigra Hagen (Oligotomidae). Upper, lateral aspect. 
Lower, ventral aspect. 



gonapophyses, they appear as a pair of rod-like struc- 
tures fused beneath the duct's apex, the orifice of 
which is microspiculate in many species. The scle- 
rotic portions of such structures are most apparent in 
cleared, slide preparations. 

The need for a well-developed intermittent or- 
gan is lessened by the fact that copulatory union is 
accomplished and prolonged by use of processes, 
lobes and hooks on the ninth, tenth and eleventh 
somites and, in many species by a clasping action of 
the left cercus and/or its basipodite. These, the pri- 
mary characters used in classification, must, how- 
ever, be used with caution because of frequency of 
convergence. For example, fusion of segments of 
the left cercus occurs in many unrelated taxa and. in 
some cases, as a variable within a species. 



External genitalia of males 

("Terminalia") 

Complex, often bewildering, male abdominal 
terminalia distinctions, often complicated by 
convergences, are fundamental characters in system- 
atic studies. 

Early in its penultimate instar a male's abdomi- 
nal apex is identical to that of other nymphal stages. 
Accordingly, the tenth tergum is triangular and un- 
modified; the eleventh (epiproct) is represented by a 
small, rudimentary sclerite (EP) just beneath the acute 
apex of the tenth; the ninth stemite is transversely 
quadrate; the anus is flanked by large, triangular, con- 
vex paraprocts (LPPT and RPPT); the cerci and their 
fleshy basipodites (LCB and RCB) are symmetrical. 
Later in the penultimate instar future changes are pre- 
saged by distortions due to developments within; the 
tenth tergum may enlarge toward its left side, the ninth 
stemite (H) may develop a small medial lobe (HP), 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



41 



laterotergites 7th tergum 8th tergum 9th tergum 



spiracle . 



10th tergum 



7th 
pleurites storm 



pleurite-- 




laterotergites ' 



Rgure 38. Abdominal terminalia of adult female Metoligotoma ingens Davis (Australembiidae). Upper, lateral aspect. 
Lower, ventral aspect. 



and its left cercus-basipodite often exhibits signs of 
profound changes. The appearance of highly modi- 
fied adult stnactures following ecdysis and reduction 
of some nymphal structures, e.g., the right paraproct, 
is most interesting. 

The terminalia usually are intricately developed 
to insure prolonged copulation and perhaps owe such 
complexity to sperm competition within the species 
to improve sexual union and thereby assurance that 
the contents of a particular male's spermatophore will 
have time to enter the spermatheca. 

In many genera the basal margins of terga of sub- 
terminal abdominal somites are extended forward as 
especially large apodemes anchoring intertergal 
muscles which elevate the terminalia during copula- 
tion. Such males often move about outside of galler- 
ies with the terminalia curled forward in the manner 
of male scorpionflies and earwigs. 



In Clothoda, the order's most plesiomorphic ge- 
nus (Frontispiece and Figs. 39—41). the terminalia 
are almost perfectly symmetrical and basically simi- 
lar to those of nymphs and females. In C. nobilis 
(Gerst.), the most plesiomorphic species of the or- 
der, the tenth tergum (10) is short, narrowly trans- 
verse, and medially uncleft. The tergite's caudal apex 
is turned upward as a thin somewhat translucent 
medial flap (MF). The medio-basal portion of the 
tergite is non-setose, weakly sclerotized, shallowly 
depressed and projected cephalad as an area I have 
termed medial sclerite (MS). This is a neutral area 
separating the more vaulted, sclerotic, setose, incipi- 
ent hemitergites (10 L and 10 R) to which the cer- 
cus-basipodite muscles are attached. 

In Clothoda longicauda Ross (Figs. 39A. 40B). 
a slightly more apomorphic species, the medial area 
of the tergite is partially membranous, forming a 
branched medial cleft which, in adult males of most 



42 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



Other species of the order, divides the tenth tergum 
into well-defined hemitergites { 10 L and 10 R), each 
of which has a distinct copulatory function. 

Viewed from the caudal aspect (Fig. 39A) it is 
apparent that the medial flap (MF) of C. longicauda 
bears a small sclerite on its ventral surface (perhaps 
a vestige of the eleventh tergite). Also visible is a 
fleshy lobe above the anus which apparently is the 
epiproct (EP), a rudiment of somite eleven. This 
conclusion is confirmed when one views the caudal 
aspect of Archembia batesi (McLachlan) (Fig. 39B) 
and notes that the ventral sclerite of the medial flap 
(MF) has elongated and is extended onto the epiproct 
which has become an extensive supra-anal pad (EP). 
A transverse fold, or hinge, occurs where the epiproct 
levels off. 

In C. longicauda. indistinct, small lobes, visible 
on either side of the medial flap (MF) appear to be 
precursors of hemitergal processes (10 LP and 10 RP). 
In Archembia batesi (Fig. 39B). these lobes are dis- 
tinct processes (10 LP and 10 RP) while the medial 
flap (MF) remains prominent. This condition pre- 
vails in many genera, as exemplified by Dihybocercus 
limaris (Navas) (Fig. 39C). In most of these genera 
the medial flap (MF) has rotated clockwise so as to 
almost parallel the longitudinal axis of the medial 
cleft. Incidentally, in Dihybocercus and other 
Embiidae, there is a small pouch at the forward end 
of the medial flap. It is likely that this end of MF 
produces glandular secretions of significance during 
copulation. This deserves investigation. However, 
in at least one major section of Embiidae, the flap 
(MF) usually is reduced to a longitudinal, sclerotic 
ridge, or it may have completely atrophied. 

In all genera of Clothodidae, except Clothoda. 
the medial flap (MF), or at least its caudal angle, 
seems to assume the function of the right process 
(10 RP), for there is no flap-like structure in the nor- 
mal posifion of the medial flap. In these genera the 
epiproct (EP) is a broad, supra-anal pad. often with 
a narrow, but prominent, longitudinal sclerite. Such 
conditions, especially that of at least portions of the 
medial flap (MF) serving as a process, are charac- 
teristic of Enveja. Oligotomidae, Teratembiidae and 
other taxa. Interestingly. Chromatoclothoda nana 
Ross is well on its way toward becoming oligotomoid 
in terminalia structure (Ross, 1987:34). 

If this tentati\e interpretation is correct, then at 
least portions of the right tergal process (10 RP) are 
analogous, not homologous. It is therefore possible 



that a major taxonomic division of the order begins 
within the family Clothodidae. It will be noted that 
the longitudinal membranous area between the me- 
dial flap (MF) and the right hemitergite (10 R) of 
most Embiidae has become transverse in 
Oligotomidae and Teratembiidae and partially to 
completely separates MF + 10 RP from an often 
much-reduced 10 R. This enables MF and 10 RP to 
hinge directly ventrad, or even forward, beneath the 
hypandrium (H) during copulation. 

Because of observation limitations within silk 
galleries, copulation is difficult to observe. However, 
in a male specimen of Aposthonia (Oligotomidae) pre- 
served in alcohol, copulatory positions of various 
structures were fixed. In this specimen the prob- 
ably-composite 10 RP (MF -I- 10 RP) is folded down 
and forward completely beneath the hypandium (H). 
The epiproct (EP) is also pulled down so that neither 
of the two structures is visible from above. Appar- 
ently contraction of strong inter-tergal muscles at- 
tached to EP is the force that moves the composite 
right tergal process (which apparently lacks muscles). 
The specimen also exhibits 10 LP pressed against the 
inner side of the basal segment of the left cercus with 
its complex apex vertical and faced to the right. The 
hypandrium process (HP), forming a rigid trough for 
the gonopophysis, is directed upward, like an erect 
human penis and is pressed against the sclerotic mar- 
gins of 10 Land 10 LP 

Throughout the order, the left hemitergite (10 L) 
is well defined, its margins usually sclerotic and in- 
flexed and its surface vaulted to provide especially 
strong anchorage for large muscles serving the im- 
portant clasper function of the left cercus, or its 
basipodite. 

The left hemitergite's process (10 LP) often con- 
sists of an inner talon and an outer flange which of- 
ten is thin, or fleshy. In many species, however, the 
outer flange is greatly reduced, or absent. The left 
hemitergite's process assumes many forms consis- 
tent within a species and thus is especially useful in 
systematic studies. It probably is the most important 
structure for providing rigid guidance of the apex of 
the endophallus into the vulva. 

Often the left hemitergite ( 10 L) is clearly sepa- 
rated from other portions of the tenth tergite by a 
submedial. membranous cleft which may extend to 
the basal margin of the tergite. In many other spe- 
cies, however, the basal portions of the cleft is ab- 
sent due to fusion of the inner-basal margin of 10 L 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 1 



43 



llthtergite^/;^ 
10 LP 



left hemitergtte 

lOL 



left 

cercus ■ bosipodite 

LCB 



medial flap MF 
eprproct EP 
^^ -10 RP 



^ right paraproct 

RPFT 




right hemitergite 
OR 



right 

cercus ■ bosipodite 

RGB 
'cercus socket 



, right hemitergite 

10R 



-/''-^H--RCB 



right paraproct 
RPPT 



left hemitergite ^ 



LCB 




right hemitergite 



ght paraproct 



left paraproct 



hypandrium process 



Figure 39. A. Caudal aspect of male terminalia of Clothoda longicaitda Ross (Clothodidae). B. Caudal aspect of 
Archembia batesi McL.) (Embiidae). C. Caudal aspect of Dihybocerciis hmaris (Navas) (Embiidae). These drawings 
show increasing complexity of the terminalia. 



44 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure 40. A. Dorsal aspect of male terminalia of Clothoda 
nobilis (Gerst.), the most plesiomorphic species of the or- 
der. B. Dorsal aspect of terminalia of C/or/joiia /ongicflMi/a 
Ross, a slightly more aposematic species of Clothodidae. 



with the medial sclerite (MS), as in most 
Teratembiidae (Figs. 52, 53). The medial sclerite 
(MS) often is obscure, or absent, but in Teratembiidae 
it is extensive, fused to the inner-base of 10 L and 
usually projects (often acute in form) beneath the 
entire left half of the ninth tergite (9). The fold of its 
left side is continuous with that of the outer side of 
the left hemitergite (10 L) and thus provides espe- 
cially strong muscle anchorage. 

The ninth stemite, or hypandrium (H), is a broad, 
quadrate, subgenital plate which usually has a weak 
basal margin but often has strong lateral margins. In 
plesiomorphic species it is symmetrically produced 
medially as a caudal process (HP) which subtends 
the apex of the endophallus. In apomorphic species 
it often is angled leftward and complexly modified, 
as in Dactylocerca Ross (Anisembiidae) (Fig. 47). 
The basal, non-setose, sclerotic portion of the para- 
procts (LPPT and RPPT) often are closely associat- 
ed, or fused, with the caudal angles of H. Indeed, the 
left paraproct (LPPT) may fuse to become the scle- 
rotic, left-caudal margin of H and, in some species 
of Teratembiidae, the hypandrium process (HP) is 
completely atrophied and the left paraproct becomes 
the sole subgenital support (Fig. 42). 

Primitively, the paraprocts (LPPT and RPPT) are 
equal in size and form. Each consists of a fleshy, 
setose, distal portion flanking the anus, and a basal, 
sclerotized, non-setose portion. In Archembia (Fig. 
39B), the distal (caudal) portion of the paraproct may 
be membranous and setose and may atrophy while 
the basal portion may fuse with adjacent structures. 
This figure also illustrates the beginning of leftward 
asymmetry of the paraprocts. 

The basal segment of the left cercus (LC,) may 
be unlobed, as in clothodids, but more often it has a 
prominent inner lobe bearing numerous, small, conate, 
peg-like setae ("echinulations") which enhance the 
segment's copulatory grip. As further improvement 
of this "tool," especially in some Anisembiidae, the 
distal segment is "absorbed" into the basal to form 
an unjointed clasper. The extreme example is in 
Dactylocerca Ross in which the segment is long, ar- 
cuate and "embraces" the females left side (Fig. 47). 
Such composite left cerci have independently devel- 
oped on many unrelated evolutionary lines. 

On a distinct evolutionary tangent, especially in 
Teratembiidae, the copulatory grip is performed by 
the extreme base of the basal segment, perhaps by 



ROSS; EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



45 



medial flap MF 
_JUh terglte EP 




right paraproct 
RPPT 

right 
cercus 



hypandrium H 
9fh sternite 



left 
cercus - basipodite LCB 



Figure 41 . Lateral aspect of terminalia of Clothoda nobilis showing upturned apex of the tenth tergite which becomes the 
medial flap (MF). and the rudimentary eleventh tergite which becomes part of the epiproct (EP). 




Figure 42. Atrophy of hypandrium process (HP) as the ventral support of the ejaculatory duct and assumption of this 
function by the left paraproct (LPPT) in males of two new genera of Teratembiidae. A. New species from Kenya with HP 
still serving as ventral support. B. New species from India. C. New species from Nigeria. Also shown in Figs. 51-52. 



46 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



the left cercus basipodite (LCB) which has varied 
mesal processes (Figs. 52, 53). 

The right cercus of males rarely is modified as it 
apparently has limited, or no function, in clasping 
the female and, in many cases, at least the basal seg- 
ment is partially or entirely desclerotized. In very 
few species, however, the inner face of the basal seg- 
ment is sclerotic and even more rarely distally in- 
wardly lobed. Such lobes never are echinulate. In 
the Australian family Australembiidae the basal seg- 
ment always is globular. 

There are probably many other factors and 
structures prolonging copulatory union. For example, 
males often grasp the female's head with highly 
modified mandibles. Dense, large setae on the 



hemitergites and inner sides of the cerci, as in 
Pachylembia Ross, may assist. The reader should 
refer to Figures 44-53, a "portfolio" of terminalia 
figures at the close of this section, which show some 
of the diversity of terminalia within the order. 

Anomalous male terminalia 

A small percentage of males have anomalous, 
"mirror-image," terminalia in which normally devel- 
oped structures are completely reversed left-to- 
right(Fig. 43A). In other anomalous specimens, struc- 
tures of the left side are symmetrically repeated on 
the right side (Fig. 43B), or those of the right are 
repeated on the left (Fig. 43C). There also are occa- 
sional bilateral gynandromorphs. These are conspic- 
uous in the case of winged species — the female side, 
of course, being wingless. 




Figure 43A. Anomalous terminalia of OUgotoma 
greeniana Enderlein of India, "mirror-image." 





Figure 43B. Anomalous Aposthonia minuscula 
(Enderlein) of India. Left side repeated on right side. 



Rgure43C. Anomalous Diradius plaumanni (Ross) of 
S. Brazil. Right side repeated on left side. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



49 




Figure 48. Metoligotoma illawarrae Davis Austral- 
embiidae). Eastern Australia. 



Figure 49. Embonycha internipta Navas (Embonychidae). 
Chapa, northern Vietnam. 



50 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure 50. Enveja bequaerti Navas. Central Africa. 



Figure 51. OUgotoma nigra Hagen (Oligotomidae). 
Middle East, introduced into southwestern USA and Aus- 
tralia. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 



51 




Figure 52. Oligeinbia capensis Ross (Teratembiidae) 
Cape Region. Baja California, Mexico. 



Figure 53. Paroligembia n. sp. (Teratembiidae). Ethio- 
pian highlands. 



52 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Literature Cited 

Alberti,V. G. and V. Storch. 1976. Transmissions und 
rasterelaktronenmikorshkopische Untersuchung 
der Spinndrusen von Embien (Embioptera, In- 
secta). Zool. Anz. (Jena) 197:179-186. 

Barlet, J. 1985. Le Pterothorax du Male d'Embia 
surcoiifi Navas (Insectes, Embiopteres). Bull. 
Soc. Royal Sci. Liege 54:349-360. 

Earth. R.. 1954. Untersuchungen an den Tarsaldrasen 
von Embolyntha batesi McLachlan. 1877 
(Embioidea) Zool. Jahrb. (Anatomic) Jena 
74:172-188, 22 figs. 

Bitsch, J. and S. Ramond. 1970. Etude du squelette 
et de la musculature prothoraciques d'Embia 
ramburi R.-K. (Insecta: Embioptera). Com- 
paraison avec le structure du prothorax des autres 
Polyneopteres et des Apterygotes. Zool. Jahrb. 
(Anatomic) 87:63-93. 12 figs. 

Carpenter, F. M. 1950. The lower Permian insects of 
Kansas. Part 10. The order Protorthoptera: the 
family Liomopteridae and its relations. Proc. 
Amer.Acad.ArtsSci. 78:185-219, 11 figs., 2 pi. 

Carpenter, P. M. 1976. The Lower Permian Insecta 
of Kansas. Part 12 Protorthoptera (continued), 
Neuroptera. additional Paleodictyoptera. and 
families of uncertain position. Psyche 83:336- 
376, 23 figs. 

Cockerell, T. D. A. 1908. Descriptions of Tertiary 
Insects n. Amer. Jour. Sci. (4) 25:227-232. 

Comstock, J. H. 1918. The Wings of Insects. 
Comstock Publishing Co., Ithaca. New York. 

Crampton. G.C. 1926. A comparison of the neck 
and prothoracic sclerites through the orders of 
insects from the standpoint of phylogeny. Trans. 
Amer. Entomol. Soc. 52:199-248, pis. 10-17. 

Davis. C. 1936. Studies in Australian Embioptera, 
part 1 Systematics, part II. Further notes on sys- 
tematics. Proc. Linnean Soc. New South Wales 
61:229-258, 50 figs., pi. XII. 

Davis, C. 1939. Taxonomic notes on the order 
Embioptera III. The genus Burmitembia 
Cockerell. Proc. Linnean Soc. New South Wales 
64:369-373. fig. 



Enderlein. G. 1909. Die Klassifikation der Embiidinen, 
nebst morphologischen und physiologischen 
Bemerkungen besonders iiber das Spinnen 
derselben. Zool. Anz. Leipzig 35:166-191. 

Enderlein. G. 1912. Embiidinen. Coll. Zool. Selys 
Long-champs, Cat. Syst. et Descr. Ease. Ill, 76 
figs, 4 plates. 

Hagen, H. A. 1861. Synopsis of the Neuroptera of 
North America, with a list of the South Ameri- 
can species. Smithsonian Misc. Coll. 4(l):xx -i- 
347 pp. (Citation as 1862 is erroneous). 

Henning, W. 1981. Insect Phylogeny, American Edi- 
tion. John Wiley and Sons, Chichester, N.Y., 
Brisbane, Toronto, xix -1-514 pp., 143 figs. 

Hong and Wang. 1993. Acta Palaeontologica Sinica 
32:141-150. 

Hurd. P D.. R. E Smith. J. W. Durham. 1962. The 
Fossilferous amber of Chiapas. Mexico. Ciencia 
(Mex.) 21:107-118. 

Imms, A. D. 1913. Contributions to a knowledge of 
the structure and biology of some Indian insects, 
I. On Embia major sp. nov. from the Himalayas. 
Trans. Linnean Soc. London (Zool.) (2)11:167- 
195, 6 figs. 3 pi. 

Kristensen. N. P. 1975. The phylogeny of hexapod 
"orders." A critical review of recent accounts. 
Zeitschrift Zool. Systematik Evolutions- 
forschung 13(1):1^M. 8 figs. 

Matsuda, R. 1960. Morphology of the pleuro-ster- 
nal region of the pterothorax in insects. Ann. 
Entomol. Soc. Amer. 53:712-731, 38 figs. 

Matsuda, R. 1965. Morphology and evolution of the 
insect head. Mem. Amer. Entomol. Institute 
53:1-334. 

Matsuda. R. 1970. Morphology and evolution of the 
insect thorax. Mem. Entomol. Soc. Canada 
76:1-431. 172 figs. (pp. 132-138, figs. 49-51). 

Matsuda, R. 1976. Morphology and evolution of the 
insect abdomen. Pergamon Press. Oxford and 
New York, viii -i- 532 pp., 155 figs. (pp. 158-160, 
fig. 34. 

Melander. A. L. 1902. Two new Embiidae. Biol. Bull. 
3:16-26. 4 figs. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 1 



53 



Mukerji. 1928. On the morphology and bionomics 
oiEmbia minor n. sp., with special reference to 
its spinning organ. Rec. Indian Mas., Calcutta 
29:253-282. 10 figs.. 1 pi, 

Pictet, F. J. 1854. Traite de Paleontogie, ou Histoire 
naturelle des animaux fossiles. Paris, J. B. 
Bailliere, 2nd ed., Vol II. 10 pis. 

Rahle. W. 1970. Untersuchungen an Kopf und Pro- 
thorax von Embia ramburi Rimsky-Korsakov 
1906 (Embioptera, Embiidae). Zool. Jahrb. 
Anatomie 87:248-330. 

Rimsky-Korsakov. M. 1914. Uber den Ban und die 
Entwicklung des Spinnapparatus bei Embien. 
Zeit. Wiss. Zool, Leipzig 108:499-519. 

Ross, E. S. 1956. A new genus of Embioptera from 
Baltic Amber. Mitt. Geol. Staatinst. Hamburg 
25:76-81. 2 figs. 

Ross, E. S. 1987. Studies in the insect order 
Embiidina: A revision of the family Clothodidae. 
Proc. California Acad. Sci. 45:9-34. 12 figs. 



Slifer, E. H. and S. S. Sekhon. 1973. Sense organs 
on the antennal Flagellum of two Species of 
Embioptera (Insecta). Jour. Morph. 139:211- 
216. I fig.. 3 pis. (pp. 218-225). 

Snodgrass, R. E. 1935. Principles of Insect Mor- 
phology. McGraw-Hill Book Co., New York and 
London, ix -i- 667 pp. 

Storozhenko. S. Y. 1997. Fossil history and phylog- 
eny of orthopteroid insects. Chapter 4 in The 
Bionomics of Grasshoppers, Katydids and their 
Kin. S. K. Gangwere. M. C. Muralirangan and 
M. Muralirangan editors. CAB International, 
Oxon, England and New York, xiii + 529 pp. 
(Chapter 4, pp. 59-82). 

Szuinik. C. A. 1994. Olifieinbia vetiista. a nevj fossi\ 
teratembiid (Embioptera) from Dominican Am- 
ber. Jour. New York Entomol. Soc. ( 102)1 :69-73. 

Zeuner, F. E. 1 936. Das erste Protoperlar aus 

Europaischem Pemi und die Abstammung der 
Embien. Jahrb. Geol. Landesanstalt 56:266-273. 



Part 2 



A Review of the Biology of Embiidina 



Summary 

The biology of Embiidina is reviewed and 
illustrated with many of my photographs. Stressed 
are the evolutionary restrictions imposed by life 
almost completely confined to self-produced, narrow, 
silk galleries. This review also covers diverse topics, 
such as: diet, locomotion, social behavior, mating, 
eggs and their protection, development, ecological 
and geographic ranges, natural enemies, and 
diseases. 

In writing EMBIA Part 1, on anatomy of the 
order, it was necessary to discuss the relationship of 
structure to function, especially in reference to the 
wings. Therefore, the reader must expect some 
repetition of information in the two parts. 

Methods 

To secure specimens for a comprehensive, world 
scope coverage of embiid taxa, 1 made many, often 
extensive, collecting trips throughout the order's 
range. For example, almost all countries of Africa 
were visited during eight excursions covering about 
four years in all. Nine months were spent in India, 
Bangladesh and Pakistan, several months in 
southeastern Asia, Australia and many months in 
significant regions of the Americas, and other places. 

To be effective, and to avoid wasting time and 
funds in hotels, major trips were made in personally- 
designed camping vehicles (see National Geographic 
articles in March 1961 and September 1965 issues). 
To get a broad representation of higher taxa, 
vegetation and life zone maps were used to determine 
routes of travel. 

Such extensive fieldwork over a fifty-year period, 
offered opportunities to observe embiid biology, but 
only briefly, for life histories are often a year in 
length. Prolonged observations had to be made in 
cultures maintained in my Academy and home 
laboratories (Fig. 43). However, because as many 
as 800 cultures resulted from a single eighteen-month 
expedition, it wasn't possible to make an in-depth 
study of any one species. 



General biology 

In spite of presumed great antiquity and isolation 
of taxa on long-separated continents, the biology of 
embiids is remarkably uniform, as it is in several 
other ancient arthropod groups, such as scorpions 
and cockroaches. In embiids, order-defining 
characters and biological uniformity result from 
perfection of evasive movement in a self-produced 
micro-environment, one which literally and 
figuratively has "channeled," or limited 
diversification of the order's anatomy and biology. 
The key factor, of course, is life almost entirely 
confined to narrow silk galleries (Figs. 1, 2 and 3). 

The galleries are produced by unique foretarsi 
swollen by numerous, perhaps hundreds, of globular, 
blastula-like, syncytial glands within the basal 
segment (Fig. 2). Viscous silk is conducted from 
each gland via a narrow duct to an opening at the 
tip of a hollow, seta-like silk-ejector. These are 
located mostly on the thin, ventral surface of the basal 
segment and to a much lesser extent on the mid- 
tarsal segment. In Oligotoma nigra Hagen, for 
example, there are approximately 150 such ejectors 
on each tarsus and thus a corresponding number of 
silk strands may simultaneously issue with each 
tarsal stroke. As both legs spin, silk webbing is 
produced with remarkable rapidity. Indeed, 
considering production speed and quantity, embiids 
may rival spiders as the most efficient silk-producing 
organisms on Earth. 

Except perhaps for their number, it is assumed 
that the glands are similar in all developmental stages 
of all species of the order. Even first instar and 
teneral individuals are able to spin and, remarkably, 
the ability continues throughout adult life. 

The galleries compose an expanding labyrinth 
usually produced and occupied by the brood of a par- 
ent female. In some species, however, it is neces- 
sary for early stage nymphs to disperse and estab- 
lish independent galleries so as to avoid injury, or 
death, due to sibling hostility. It should be noted, 
however, that such hazards are likely to be intensi- 
fied in crowded laboratory cultures in which most of 
my observations were made. 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure I. Adult Parwhagadochir birabeni (Navas) (Einbiidcii.") ot Argentina, showing typical ap- 
pearance and posture of all female embiids. Pale bands between thoracic somites characterize many 
species of the order. 




Figure 2. Adult female of Ptilocerembia n. sp. (Notoligotomidae) from Malaya exhibits universal 
spinning foretarsi, large hind femora and short, Uvo segmented cerci. The \Tjlva opens between the 
dark, sclerotic eiahth and ninth abdominal sterna. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 2 




Figure 3. Nymphs radiate in galleries, often used in common, which fit their body size. A 
portion of the parent's larger gallery crosses the top of this photograph, Donaconethis n. sp. 
(Embiidae), Eritrea. 



Embiids are highly thigmolactic and produce gal- 
leries just narrow enough to maintain constant con- 
tact of body vestiture with walls. Advantageous 
confmement within galleries apparently has governed 
anatomical and behavioral trends over a long period 
of evolutionary time. Thus, the spinning organs ap- 
pear to antedate and to have regulated the evolution 
of all order-defining characters. Such characters must 
have become fixed in Upper Paleozoic or Lower Me- 
sozoic times and fully distributed within the order 
before the breakup of Pangaea. 

Gallery diaineter correlates with the size of the 
embiid frequenting the particular section of a colony. 
In most species first and second instar nymphs, after 
clustering for several days near their mother (Fig. 4) 
commence to radiate outward in their own galleries 
which are increased in diameter and extent as the 
maker grows. Because offspring of most species hatch 
from a single egg mass and develop in unison, the 
galleries of a brood usually becoine interconnected 
and used in common. 

In addition to the tarsal silk glands, ordinal char- 
acters include an elongate, supple body; a prognathus 




ficil Hi 4 1 nsl and see 
ter near the parent fema 



'lid IllsKll n\ 1 

e, Eiiibiu sp. 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 







Figures 5 and 6. Especially in and regions, parent females often find refijge in rock crevices. Follow- 
ing rains, galleries of the brood radiate on the rock's surface as lichens are "grazed."" During excessive 
heat, or bush fires, the embiids crowd back into the crevice, Scelembia n, sp,, (Embiidae), Angola. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



head with a sclerotic, ventral bridge; short legs, effi- 
cient backward movement; equal wings with reduced 
vein strength except for sinus veins which temporari- 
ly can be stiffened with blood pressure; tactile, two- 
segmented cerci; universal neoteny in females, 
partially or entirely so in males of some species. 

These, and other features treated elsewhere in this 
work, maximize survival almost entirely dependent 
on confinement in narrow galleries. Such galleries 
are constantly extended within or on edible surfaces. 
Therefore, embiids can feed without leaving their 
galleries. Food is very simple. All species are pri- 
marily phyto-scavengers but some may also occasion- 
ally eat live mosses, lichens, etc. There are no 
predatory species. 

Some researchers have suggested that the galler- 
ies control humidity, others, including me, have con- 
cluded that their primary function is protection from 
predators and parasites. Although it seems unlikely 
that the galleries can regulate humidity to any appre- 
ciable extent, I have observed the dense gallery walls 
are at least temporarily impervious to water and thus 
may protect the occupants, especially soil-inhabiting 
species, from short-temi flooding or habitat satura- 
tion after heavy rainfall. 

Embiids, such as Haploembia spp., inhabiting 
regions with cold periods, even those temporarily 
blanketed in snow, completely enclose themselves in 
a cocoon spun within the galleries. I have noted them 
in the introduced species, Haploembia solieri (Ram- 
bur), in California, as well as in Mediterranean and 
Turkish regions. The habit may be widespread in 
species living in seasonally-cold regions. As with 
cocoons of moths, and other insects, the enclosures 
must function primarily as predator barriers. This is 
especially important to embiids in seasonally-cold 
regions because potential enemies, such as predaceous 
beetles, may be able to hunt prey at low temperatures 
which immobilize embiids. 

The primary advantage of gallery life seems to be 
predator-avoidance and this is increased when gal- 
leries extend beneath or within, solid objects (Figs. 
5, 6). Protection in exposed galleries may be some- 
what indirect because the silk isn't strong enough to 
wall off most predators. A predator's initial contact 
with a web surface probably broadcasts a tactile warn- 
ing which stimulates rapid, usually reverse, move- 
ment into a deeper, more rigid recess within the 
labyrinth. Edgerly's study (1997) of ant entrance 



holes in gallery walls, based on Antipaluiia urichi in 
Trinidad, appears to contradict this conclusion, at 
least in her study area. While collecting embiids 
throughout the order's geographic and ecological 
range, I have frequently encountered embiid colo- 
nies in direct contact with those of ants, often under 
the same stone, without any apparent molestation by 
the ants. Ants, however, are probably the principal 
predators of embiids whenever they leave their gal- 
leries. 

The value of rapid silk web production was ap- 
preciated by me when, experimentally, I released em- 
biids on tropical tree trunks. An exposed embiid 
immediately retreats into the nearest bark crevice and 
at once begins to cover it with a silk web. This barri- 
er is steadily improved and extended and, if favor- 
ably located, may become the locus of a new labyrinth. 
It is more likely, however, that the exposed embiid 
will immediately be seized by an ant or other preda- 
tor before it can get into a crevice and spin a barrier. 
It may be said that an embiid outside of its gallery is 
almost as much out of its element as a fish out of 
water. 

Life in silk galleries may offer other benefits. It 
is probable that embiids conserve energy by having 
pre-constructed, smooth-surfaced runways to and 
from a food source. Silk itself may be an excretory 
by-product put to good physiological use. There is 
also the advantage of being able to silk-partition fe- 
cal pellets from within the galleries, thereby main- 
taining debris-free avenues of movement. After 
excreting a pellet, an embiid snips a hole in the adja- 
cent gallery wall and, using its mandibles as tongs, 
places the pellet outside and then closes the opening 
with silk. Fecal pellets accumulated between galler- 
ies tend to strengthen their walls (Fig. 7A) and, when 
deliberately placed atop a labyrinth's surface on a 
tropical tree tmnk, serve as a medium for growth of a 
microflora enhancing the cover. The habit of pro- 
tecting, or not protecting, exposed galleries with a 
surface covering may characterize a species, or even 
a genus. In addition to fecal pellets, some species 
pulverize outer bark and cause gallery surfaces to 
become dusted with chewed powdery debris which 
may almost completely conceal the colony (Figs. 8, 
9). In contrast, species of many other genera ne\er 
cover their galleries and thus they can be seen from a 
considerable distance (Figs. 25, 26). 

Another type of feces disposal invoh-es their ac- 
cumulation in low mounds here and there on the la- 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 







:^.v 



Figure 7A. Embiids place their dry fecal pellets outside 
of their galleries, thereby insuring debris-fi^ee avenues of 
movement and a strengthening of gallen,' walls. Archembia 
batesi (McL.) (Embiidae) m an Amazon rainforest. Sur- 
face layer of silk removed. 




Figure 8. In contrast to Archembia, many embiids, as a 
generic habit, deposit feces on exposed surfaces of their 
galleries. Chromatodothoda n. sp., ( Clothodidae). Ec- 
uadorian montafia. 





Figure 7B. Galleries of Archembia batesi are conspicu- 
ously white and, in this case, aligned with bark cre\ices. 



Figure 9. Galleries of a new genus and species from south- 
eastern Asia's Golden Triangle, are completely concealed 
beneath pulverized bark particles and feces. Doi Pue, Thai- 
land. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 2 



byrinth's substrate. These are then progressively 
covered with layers of silk and appear as low white 
mounds similar to those covering egg masses of some 
species (Fig. 42). 

In tropical forest habitats another type of protec- 
tion involves prolonged use of old galleries as a sur- 
face cover over those spun beneath by successive 
broods. Thus they can function somewhat as a layer 
of bark. Galleries themselves may also have pro- 
longed use and consequently silk becomes increas- 



walls through flexibility (vein atrophy) and alternat- 
ing stiffening for flight by blood pressure in sinus 
veins. The ultimate accommodation is wing size re- 
duction, or complete loss, through neoteny (or pae- 
domoiphosis). 

These wing specializations are so complex and 
universal that it is inconceivable that they evolved 
solely to increase survivability of adult males. In- 
stead, they must have developed ages ago when the 
order was confined to tropical zones of Pangaea as a 




Figure 10. Adult male tlunng del'ensivc backward movement. Temporarily flexible wings bend forward, thereby reduc- 
ing friction which could slow escape. Actually, wings when not used in flight can bend at any point — even crumple. New 
genus and species of Oligotomidae, Thailand. 



ingly dense and obscured by surface debris. Thus, in 
Amazonia, galleries of species of Clvomatoclolho- 
da Ross (Fig. 8), in contrast to the conspicuous, white 
galleries of more plesiomorphic Clothoda Enderlein 
(Fig. 23), both family Clothodidae, can be located 
only by random tweezer-scraping of likely surfaces, 
such as the underside of laterally projected branches 
or ledge overhangs. 

The presence or absence of wings and their pecu- 
liarities are directly related to predator-avoiding re- 
verse movement in silk galleries — the need to reduce 
or overcome wing friction, or snag, against gallery 



means of increasing predator-avoidance by adult fe- 
males which must live long enough to produce and 
guard eggs and early instar broods. In conti-ast. males 
are short-lived and contribute only sperm to the re- 
productive process. Later, however, more effective 
reverse locomotion of females was achieved through 
complete apterism by means of neoteny. In a sense, 
females of all extant species are second or third in- 
star reproductive nymphs increased in size. Although 
males of most species possess wings, there is a ten- 
dency on almost all evolutionaiy lines for them to 
become completely apterous or brachypterous 
through neoteny. 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Embiid wings are non-deciduous, nearly equal in 
size and shape, wide spaced in thoracic attachment, 
flat and unfolded when in repose over the dorsum, 
have pigmented stripes separated by hyaline inter- 
vals following the courses of all longitudinal veins, 
and, most important — indeed unique — some veins are 
broad, glossy, cuticularized, longitudinal, blood si- 
nuses. Except for slight anal ( vannal) lobing in wings 
of certain species of the rather plesiomorphic genus 
Archembia Ross, the anal area of embiid wings is 
greatly reduced. Such reduction even occurs in 
Clothoda Enderlein, the order's most plesiomorphic 
genus. It is probable, however, that embiids had an- 
cient ancestors with a well-developed anal area in the 
hindwings comparable to that of Plecoptera andMas- 
totermes in Isoptera. 

Thus the principal evolutionary trend in embiid 
wings wasn't improvement for flight but, instead, to- 
ward rendering their possession less of a handicap 
during movement within galleries. The ultimate ac- 
commodation is complete aptery of all females as well 
as of males of many taxa. A similar disadvantage of 
wing possession was faced by sexual termites adapt- 
ing toward easier movement in galleries in earth and 
wood. In this case, however, the disadvantage is elim- 
inated by wing break-off by nuptial adults prior to 
copulation. Also, lifelong aptery of most individuals 
in a termite colony is caused by endocrinal retarda- 
tion of the appearance of adult structures and func- 
tions (neoteny). A similar retardation is probably 
responsible for the universal aptery of adult female 
embiids, as well as varying degrees of brachyptery 
to complete aptery in males of many embiid species, 
genera, and even an entire family (Australembiidae). 

The trend toward apterism in males is presently 
active and has been so for at least the entire Tertiary 
period, as evidenced by complete apterism of males 
of Electroembia antiqiia (Pictet), Baltic Amber 
(Eocene?). Degrees of male apter>- and brachyptery 
occuronmostevolutionary lines, such as: (I) males 
with robust (nymphoid) bodies and short wings; (2) 
males with wing pads similar to those of various on- 
togenic stages from buds (gemmae) to fiill pads; (3) 
complete aptery without even traces of wing buds. 
In some species, such as Anisembia texana (Mel.), a 
percentage of adult males in a population have nor- 
mal wings, but most possess only tiny wing buds. In 
Oklahoma, at the northern range of A. texana. all 
males are completely apterous without traces of wing 
buds. 



Because females are universally apterous, flight 
of males cannot increase geographic range of a spe- 
cies, or enable a population to move away from envi- 
ronments uninhabitable as a result of sudden or 
gradual adverse ecological changes. Range exten- 
sion and relocation can be effected only by females 
surviving hazardous movement afoot outside of their 
protective galleries, or by being carried in materials 
transported by wind, water or human commerce. Male 
flight, however, fosters random mating and thereby 
reduces potentially disadvantageous incestuous mat- 
ings so likely in gregarious, subsocial populations. 
The complex subject of embiid wings is more frilly 
treated in my review of the order's anatomy (EM- 
BIAPart 1). 

Social behavior 

A typical embiid colony is a "gynopaedium" — a 
parent female and her brood living together. Often 
galleries of broods of adjacent females become in- 
terconnected and the nymphs intermingle without 
hostility. 

Although there is no evidence of a division of la- 
bor, or castes, some social advantages could result 
from utilization, by some species, of preexisting gal- 
leries produced by previous generations which had 
occupied the same bark surface. In rainforests a mat 
of such galleries may thus serve like a layer of bark 
protecting new galleries spun beneath by succeeding 
generations of nymphs. However, to reach ungrazed 
edible surfaces, most species produce new labyrinths 
radiating out from such initial coverings. Incidental- 
ly, the silk of new galleries of some species is laven- 
der in color. 

The most important social activity is guarding 
eggs and young by parent females in a manner simi- 
lar to that of Dermaptera (Figs. 18,42)(EdgerIy, 1987a 
and b; 1988). Early instar nymphs usually congre- 
gate near their mother and perhaps benefit from her 
presence for at least two instars (Fig. 4). As I service 
laboratory culmres, young are often inadvertently dis- 
associated from their parent but, in a short time man- 
age to reassemble in spite of the disadvantage of hav- 
ing to spin new galleries to do so. An aggregation 
pheromone may be involved in this. 

Food provisioning in arid regions by subterra- 
nean species can also be regarded as social activity 
for it tends to insure dependable, readily accessible, 
food for the brood. It also avoids energy loss and 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



risks consequential to making repeated forays to an 
often inclement and hazardous surface environment. 

Embiids may be termed subsocial communal in- 
sects. In this category females remain with their off- 
spring for a period of time and members of the same 
generation use the same composite nest without co- 
operating in brood care. Such an interaction is an 
inevitable consequence of gallery life. A lone female 
simply lays eggs in a favorable place — often within 
existing galleries, or in a new site, and her offspring 
having no need to disperse, extend individual galler- 
ies no farther than needed to reach food. Upon matu- 
rity, her offpspring do not need to disperse to find 
food or a mate. Therefore, females are likely to de- 
posit eggs not far from their own place of origin. 

In communal species, especially those occupying 
tree trunks, or rock and road bank surfaces in the trop- 
ics, a colony may grow to great size and even enve- 
lope a huge tree trunk (Fig. 26). Theoretically, such 
growth is radial — expanding at the periphery as edi- 
ble surfaces are sought. In jar cultures and under 
other artificial conditions, colony growth is three-di- 
mensional, as reported by Friederichs (1913) in ref- 
erence to an outbreak of Aposthonia gurneyi 
(Froggatt) in a sugar refinery in Australia, or of Olig- 
otoma saimdersii (Westwood) in piles of stored pea- 
nuts in Senegal. 

There are exceptions to a gregarious habit, how- 
ever. In some species, especially those found in the 
savanna woodlands of central Africa (e.g., Dinembia 
Davis spp. ), nymphs are intolerant to one another and 
must disperse soon after hatching and develop in in- 
dividual galleries to avoid injury. 

It is conceivable that if ever a form of reciprocal 
or proctodeal feeding, or body licking should evolve 
which would permit transmission of maturity-inhib- 
iting pheromones, a worker caste might develop in 
embiids. There is, however, absolutely no indication 
or need of such behavior in Embiidina. Food ex- 
change between embiids has never been observed and 
proctodaeal feeding potentials are lessened because 
excrement consists of dry pellets which are deliber- 
ately placed or partitioned outside of the galleries im- 
mediately following defecation. Nymphs hatching 
from eggs don't even eat the hardened pulverized ma- 
terial, which in part may be fecal, placed around the 
eggs by females of most species. Covering and side- 
by-side placement of eggs appears to reduce ovipo- 
sition by parasitoid wasps (Figs. 19, 42). 



Female embiids could be likened to nymphoid re- 
productives in termites, but it is most likely that their 
neotenization is programmed by regulation of juve- 
nile hormone rather than exchanges between individ- 
uals. 

Behavior of adult males 

Upon maturity an adult male usually remains in- 
active for a few days in the gallery section where fi- 
nal ecdysis took place (Fig. 12). During this time its 
derm hardens and becomes fully melanized. or pig- 
mented. Concurrently, its nymphal pelt slowly pass- 
es through the gut and is excreted. Later a male may 
wander about within the galleries and may mate with 
a receptive female, perhaps a sister, particularly in 
species with apterous males which are more likely to 
remain in a colony with sisters. Sister-mating is like- 
ly in laboratory cultures and in gregarious species but. 
in some species, males and females develop in sepa- 
rate galleries. In such cases a male must vacate his 
"personal" gallery, locate and bite his way into one 
occupied by a female (Fig. 15). 

In cultures, usually during warm afternoons, adults 
of both sexes often move to uppermost levels of a 
culture, protrude their forebodies from gallery open- 
ings (Fig. 14) and, with their heads often hypogna- 
thously angled, they rapidly vibrate their antennae. 
It is assumed that such activity encourages sexual 
contact. No investigations have been made to deter- 
mine if "calling" pheromones are released during 
such exposure, but this is likely. 

Under certain meteorological conditions, often 
just after the first rains ending a dry season, aduh 
males usually leave their galleries and take flight (Fig. 
1 1 ), or, if apterous, simply run about on the ground. 
In arid regions males of pale species have large, 
coarsely-faceted eyes and tend to fly during warm, 
humid nights when many other insects, notably nup- 
tial termites, also fly. During such periods collectors 
should always search for male embiids attracted to 
lights, not only to collect specimens, but also as a 
means of determining what nocmmal species of the 
order have colonies in the vicinity. At least a portion 
of a light sheet should be on the ground inasmuch as 
males of some pale, nocturnal species are apterous 
and can only run to lights. In some species, such as 
Aposthonia tillyardi (Davis) of western Australia, 
apterous and alate males may occur in a single local- 
ity. 



10 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Color of adult males 

Adult males of most species are melanized, at 
times with a reddish prothorax, and/or a golden 
pterothorax and metallic blue wing veins and sheen 
(Fig. 13). Dispersal of colorful males is diurnal and 
one may see alate males of such species in flight, or 
resting on vegetation but, more likely, yet rarely, they 
will be collected by random sweeping. Probably 
mortality of diumally dispersing males is very high 
due to increased exposure to birds, or the elements. 



Some protection against predation may involve Ba- 
tesian mimetic resemblance to chemically-protected 
diurnal beetles and stinging ants. 

In various, usually unrelated taxa, the distal an- 
tennal segments may be abruptly white, as also are 
one or both cercus segments. In many species pale 
intersomatic thoracic bands are present, as well as 
longitudinal, pale, pleural abdominal stripes. All such 
characteristics are due to white fat visible through a 
transparent integument. 




FiGURh 1 1 . .Aien poslurc ol Lin .iduil nvdii: about lo lake llit;tiL 1 lie strongly CUticular- 
ized wing veins of this genus are prominent in this photograph. Enveja bequaerli Navas. 
Katanga, central Africa. 




Figure 12. This teneral male will remain for a long time in one place until fully hardened. 
Incidentally, it has especially large wings, a characteristic of high altitude species. 
Pararhagadochir n. sp., (Embiidae). Machu Picchu, Peru. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 




Figure 13. Adult female and penultimate male of new 
genus and species of Oligotomidae from Thailand. The 
prothorax of both sexes of this dark brown species is bright 
orange. Body length of female 1 8 mm. 




Figure 14. Epigamic females may protmde their forebodies 
from galleries, possibly emittmg pheromones to attract 
males. '"£';)z/)/a"«/rco!(/?Navas, (Embiidae). Mozambique. 



Mating 

When a male locates a gallery containing a re- 
ceptive female, he bites an opening (Fig. 15), enters, 
and approaches the female head-on, rapidly jerking 
his body and vibrating his antennae. If the female is 
unreceptive, her reaction, perhaps varying according 
to species, may be antagonistic and dangerous. In 
some encounters a female may attempt to eat, or at 
least bite an approaching male. She lunges toward 
the male with the same motions used in defending 
eggs, or young brood. Often there is antagonism or 
fighting between males. 

When a female is receptive, there is mutual quiv- 
ering of the antennae, head and prothorax and, alter- 
nately, forward darting and retreat. Inasmuch as there 
may remain a threat to the male, males of many spe- 
cies reduce danger by grasping the fore portion of 
the female's head with the mandibles. Mandibles of 
males exhibit varying degrees of specialization for 
such a grasp, the most extreme of which are the large, 
elongate, arcuate type characterizing some species of 
Enveja Navas. Mandibles of males may also be used 
to gently nibble the female s body. Because the man- 
dibles of adult males are not used for grinding food, 




Figure 15. Adult male (new genus and species, Oligotomi- 
dae) biting an entrance into a gallery presumably occupied 
by a receptive female. Thailand. 



but primarily seem to be secondary sexual organs, . 
they have greatly diversified and are thus useftil char- 
acters in systematics. The short, robust, nymphoid 
mandibles of adult females, used for food-grinding, 
vary little throughout the order. 



12 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



Usually, a male grips a female's head across the 
frontal region but there are variations. In one abnor- 
mal case, I observed a male holding the dorsal cervi- 
cal region instead of the head. Once secure in a 
mandibular grip, the female's head usually is pulled 
to the right and the tip of the male's abdomen probes 
down her right side, thence leftward and upward be- 
neath her genital opening. It is remarkable that a right- 
side approach is constant in all species of the order. 
In any event, it explains the almost universal leftward 
asymmetry of the male terminalia (to extend its 
"reach") and why the specialized left cercus functions 
in many species as a clasper against the female's left 
side 

Because mating occurs obscurely within galler- 
ies, it is difficult to observe. However, Stefani (1953a, 
c) made detailed descriptions of copulation in Emhia 
rmnhurt R. K., Cleomia guareschii Stefani, and Hap- 
loembia i-o/Zer/ (Rambur). Earlier, Friederichs ( 1 934) 




Figure 16. Copulating Pachylembia chapalae Ross of 
western Mexico preserved in alcohol. As generic charac- 
ters, the male (left) lacks lobing on the left cercus and has 
a greatly reduced left tergal process. The copulatorv' grip 
appears to be increased by pressure of the dense, bristle- 
like setae borne on the left and right hemitergites. 



made brief observations of Embia ramburi and 
Oligotoma nigra Hagen. 

A few of my observations are recorded as fol- 
lows: 

(I) ^rcAe/H/j/a /flcomiiea Ross (Embiidae). Bra- 
zil: Rio de Janeiro. A male was observed rubbing 
his submentum against the vertex of a female's head. 
In this species, and congeners, there are relatively 
dense, often parallel, setal clumps, as well as foveae 
on the submentum in and around which white secre- 
tions collect. During this rubbing the male's anten- 
nae extended on either side of the female's body. She 
wriggled sinuously and continued a limited spinning 
movement of her forelegs. Females of Archembia 
Ross, and related genera, have a transverse, pale, of- 
ten golden, eye-to-eye band above the brain. One 
should investigate the function of this pale macula- 
tion. Is it associated with mating, or does it have a 
light-perception ftinction? 

(2) Machadoembia Ross, n. sp., (Embiidae) 
Angola: near Quilenda. 

Male grasped female's head (face to face) with 
his mandibles across her clypeus. The female fre- 
quently lurched but the male maintained his grip for 
at least a minute. During this time the male's genita- 
lia united with those of the female. These discon- 
nected before the male released his grip on the head. 
When freed the female walked off unharmed by the 
male's mandibular grip. 

During other matings. males of this species 
grasped heads of females ft-om several frontal angles, 
as well as the cervix. The female's head was twisted 
to the right as the male's genitalia quickly sought con- 
tact. In spite of much tugging, females seemed re- 
ceptive to mating. One copulation lasted about 60 
seconds. 

(3) Pflre/^fe/fl /7;a/o/-(Imms), (Embiidae). India: 
Mussourie U.R 

Male gripped female's head frontally. 

(4) Embia n. sp., (Embiidae). Ethiopia: Naza- 
reth. 

Male faced female and gripped her head with his 
mandibles behind her eyes. The male twisted so that 
most of his body paralleled the right side of the fe- 
male and the tip of his abdomen crossed beneath the 
female's genitalia. Actual genital union wasn't ob- 
served. 



ROSS: EMBIA, BIOS YSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



13 



(3) Einbia iuaurita?iicaLucas (Embiidm). Alge- 
ria: 27 mi. N. M'Sila. 

Ventral concavity of median flap (MF) covered 
inner curvature of left tergal process (10 LP). Pro- 
cess of left paraproct (LPPT) pressed into dorsal de- 
pression of left cercus lobe. Ventral nodule of LPPT 
prevented left cercus lobe from moving ventrad. 

(6) "Parembia" dobhali Ross. {Embudas). India: 
DehraDun. 

Female began to eat copulating male while the 
genitalia were still joined. 

(7) "Parembia" n. sp., (Embiidae). India: Bad- 
amtan Forest Res. W. Bengal. 

A male attempted to mate with another male 
while holding its head in his jaws (behind the eyes) 
and pushing his genitalia beneath the other male's 
wings, apparently mistaking this surface for that of 
a female's abdominal venter. Such abnormal ap- 
proaches were observed on several occasions. 

(8) Enveja bequaerti Navas, Zaire: 12 mi. S. 
Sampwe. 

Grip of head with jaws not observed but the re- 
markably large mandibles suggest that they are 
adapted for head-clasping. During one mating the 
sexes remained parallel; the male on the right side 
of the female with his terminalia angled leftward 
and upward to join the female's genitalia. They re- 
mained united for about thirty minutes. Close ex- 
amination revealed that the male's right tergal process 
was folded ventrad against the surface of its hypan- 
drium and pressed against the female's second valvi- 
fer. After the pair separated, a hard, irregular, 
gelatinous object, probably a spermatophore, pro- 
truded from the vulva. 

(9) Dactylocerca Ross, n. sp. (Anisembiidae), 
Mexico: Alamos, Sonora. 

The male's long, arcuate, one-segmented, left cer- 
cus embraced the left side of the female's abdominal 
apex from beneath. The grip was so tight that mem- 
branes at the base of the female's rigli cercus be- 
came distended. At no time did the heads connect 
although this might have occurred before the obser- 
vation began. Males of the genus have very small 
mandibles. 

(10) Australembia nodosa (Davis) (Australem- 
biidae). Queensland: Millstream Falls. Mating 
fixed in alcohol. 



Male terminalia centered beneath female. The 
apex of LCi+2 depressed membranes between left 
basal comer of H and between caudal tips of pleu- 
rite and laterotergite of somite 8. The hypandrium 
process (HP) and tenth tergite (10 RP) pushed into 
vulva. After separation, a spermatophore wasn't 
visible in the vulva. Perhaps the pair was killed in 
alcohol before copulation was completed. 

(11) New genus and species (Teratembiidae), 
Transvaal: 18 mi. S. Louis Trichardt. 

Mates faced the same direction. The male did 
not grip female's head and was somewhat beneath 
her. his terminalia turned upward to the vulva, well 
centered. Because it is unlobed, the left cercus did 
not seem to be used as a clasper. 

Later the female walked forward causing the two 
insects to face opposite directions. The female con- 
tinued to walk out of sight into a gallery dragging 
the male backward as the genital union continued. 

(12) Aposthonia Krauss, n. sp. (Oligotomidae). 
Queensland: Brookdale (coastal plain). 

In a male specimen preserved in alcohol, the 
right tergal process (10 RP) was folded down and 
completely pressed against his hypandrium (H). The 
epiproct (HP) was also pulled down and provided 
musculature for movement of 10 RP. The left tergal 
process (10 LP) paralleled the inner face of the bas- 
al segment of the left cercus (LCi) with its complex 
apex vertical, its dorsal surface facing toward right; 
its left side was appressed on the inner apex of LCi. 
The hypandrium process (HP) and the gonopophys- 
is were projected dorsad (almost vertical, like an erect 
penis) and, pressed by the inner angle of the sclerot- 
ic left hemitergite (10 L) and its process, provided 
rigid enclosure for the ejaculatory duct. 

These "mechanics" are probably universal in the 
Oligotomidae, Teratembiidae, and other taxa which, 
by convergence, have a transverse, membranous sep- 
aration of EP and 10 RP from 10 R which serves as 
a hinge permitting downward movement of 10 RP. 

Eggs and their protection 

Eggs of all species are remarkably similar (Fig. 
17). They are tubular in form, basally rounded, 
slightly curved, and have a large, slanted, strongly- 
rimmed operculum. Their general appearance is sim- 
ilar to that of bedbug eggs. 



14 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Eggs issue from the vulva with the operculum in- 
ward and are deposited within the galleries, usually 
attached to a substrate. However, in some species 
they are loosely clustered in the galleries and not 
imbedded in a hardened paste. In some species un- 




FiGURE 17. All embiid eggs are similar in shape, have a 
rimmed operculum, and are laid on their back. Haploem- 
bia solieri (Rambur) (Oligotomidae) endemic to the Med- 
iterranean region. 



covered eggs may form a tunnel within which the 
guarding female rests. Most often, however, eggs are 
laid in a single-layered cluster and are imbedded in a 
hardened paste of habitat material pulverized by the 
female and deliberately placed as the eggs are laid 
(Figs. 19, 20). It is probable that fecal pellets are 
also pulverized for this purpose. I have counted more 
than two hundred eggs in a cluster laid over a period 
of several days in one species, but the numbers may 
be much less in other species. Egg size may be con- 
stant regardless of the size of the female, i.e., those 
laid in small numbers by minute species of Oligem- 
bia appear to be as large as those laid by relatively 
huge Antipahiria females. 

The tightly clumped eggs are slightly slanted with 
the opercula exposed. Many species spin a dense 
covering of silk over the mass. Obviously, such cov- 
erings, and parental guarding, reduce the percentage 
of eggs parasitized by scelionid wasps. It is likely 
that the habit of covering eggs is related to the geo- 
graphic occurrence of the wasps. For example, in 
the Mediterranean region where such wasps appar- 
ently do not occur, the eggs oi Haploembia spp. and 
Embia spp. are uncovered and loosely clumped. 

Species within unrelated Amazonian genera en- 
close eggs in a sawdust-like matrix of chewed habitat 
particles which is densely covered with silk, thus 
forming a low mound on which the female rests (Fig. 
42), ready to challenge approaching parasites and 



K?:-.^'' 




Figure 18. As females guard their eggs, they lunge toward enemies — particularly egg para- 
sites. £>/HeH!/j/a sp. (Embiidae). Northern Zambia. 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



15 



predators. However, such protection, like most de- 
fenses in nature, is imperfect. For example, I found 
within a mass containing 5 1 eggs, 12 fully developed 
scelionid wasps clearly visible through transparent 
egg shells. 

Additional information on maternal protection of 
eggs and young is provided by Edgerly (1987a, b; 
1988, 1994) in her detailed study of plesiomorphic 
Antipaluria iiiichi (Saussure) (Clothodidae) in 
Trinidad. 




i ■>. w 



Figure 19. Most embiids, such as Antipaluria Enderlein 
(Clothodidae). reduce oviposition of wasp egg parasites 
by packing a paste of pulverized material around their eggs. 
Venezuela. 




Development 

Adult female embiids exhibit little change in ap- 
pearance from first instar nymphs except, of course, 
for increased size and coloration. Ventrally, the eighth 
and ninth abdominal paragenital stemites adjacent to 
the vulva's opening are modified, as is, of course, 
maturation of internal reproductive organs. Neoten- 
ic apterous males usually are similarly nymphoid but, 
as adults, are more melanic, or pigmented, and have 
distinct cranial and abdominal terminalia characters. 

Males destined to have wings show the first ex- 
ternal evidences during an early nymphal instar. At 
first they are merely very slight extensions of the pos- 
terior angles of the meso- and metascuta (Fig. 21 A). 
These are accompanied by increased development of 
certain setae near, and on, the lateral margins of the 
nota. The enlarged angles somewhat increase in size 
during the stadium. 




Figure 20. As eggs are laid on the sides of culture jars, 
their number and imbedding sequence can be observed. 
"Embia " surcouji Navas, Mozambique. 



Figure 21. Wing development of a typical male embiid. 
Oligotoma nigra Hagen. 



In the next instar definite wing pads appear (Fig. 
21B). Those of the mesothorax nearly reach the an- 
terior margin of the metanotum. The lateral notal 
setae now have increased in number and mark cours- 
es of future wing veins. Development of the special 
radius blood sinus vein (RBS) is indicated in some 
specimens by faint "fleshy," reddish lines which bor- 
der RBS in the adult wing. As in the previous instar, 
there is pad enlargement during the stadium. 



16 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



In the next (penultimate) instar (fifth?) (Figs. 13, 
2 1 C , 22), the wing pads are well developed and great- 
ly elongated; those of the mesothorax overlap most 
of the metascutum, the pads of which extend caudal- 
ly over most of the second abdominal scutum. They 
are broadly attached to their respective scuta. How- 
ever, these lines of attachment do not represent de- 
finitive wing bases for they are actually part of the 
posterior wing margin. 

Definitive veins are indicated by their setae and 
RJBS by even greater pigmentation of its marginal 
bands. Tracheae follow the same courses as the set- 
ae. It is significant that venation of the pads con- 
forms to that of the adult; MA being unbranched, for 
example, in all oligotomid and anisembiid wings. The 
observafion by Melander (1903) that MA (his R4+5) 
in Anisembia texana, as evidenced by the trache- 
ation in the pads, is forked in the nymph and not in 
the adult was probably an error, or anomalous. 

Expansion of wings 

During most of the penultimate nymph stage of 
males destined to be fully alate, the wing pads are 
clear, thin, and fiat. Later they become thick, opaque, 
cream white in color, and the dark lines bordering the 
radial blood sinuses are conspicuous (Fig. 22). Inci- 



dentally, at this time divisions of the tenth abdominal 
tergite are visible through the derm and the adult cer- 
ci are withdrawing basad. Finally, the nymph ceases 
movement and in a short time it is possible to observe 
ecdysis, and wing expansion. The following is my 
account of these events in a species of Embiidae from 
Africa's Ruwenzori Mountains. 

2:40 PM Nymph emerged from its last nymphal 
exuvium. The unexpanded wings at first were slen- 
der, strongly convex, thick, and cream white in color. 

2:50 PM Starting from the costal and basal mar- 
gins, the wings began to fiatten and expand. Period- 
ically, the soft adult wriggled and rotated its body. 
The abdominal terminalia were distended, all struc- 
tures were swollen, and the cerci projected laterad at 
45°. 

3:00 PM Entire basal half of the wings now broad- 
ened and flattened. The distal half remained as small 
and as narrow and fleshy as at 2:40 PM. At 3: 10 PM 
the entire costal margin had expanded with only the 
apical end of the hind margin remaining fleshy. This 
condition prevailed until 3:30 PM. 

3:50 PM Left pair of wings now completely ex- 




FiGURE 22. Late penultimate instar of a male. Note that the wing pads have thickened and that cerci 
of the adult are withdrawing fi-om the nymphal skin. Antipaluria caribbeana Ross. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 2 



17 



panded, right pair remain fleshy ventrally at apices. 
Wing veins paler than membranes. 

4: 10 PM All wings completely expanded, defin- 
itive in form and thickness. Hyaline stripes had ap- 
peared. 

9:40 PM Wings now gray in tone. Body and leg 
pigment developing. 

7:30 AM (next day) Wings now smoke black. 
Exuviae eaten during the night. 

1 :00 PM Male remains in same place in gallery. 
Much darker in color. 

9:00 AM (third day) Male still in same position. 
Darker in color 

The above appears to typify wing expansion of 
all Embiidina. Sequences have been observed on 
several occasions in distantly related species. How- 
ever, a number of variations were noted, as follows: 
In an embiid from Takoradi, Ghana, a male nymph 
about to transform to an adult, was isolated at 2 PM. 
As late as 8 PM it continued to spin silk even though 
subdermal adult structures were well advanced and 
in spite of the fact that it was destined to shed its 
nymphal skin within two hours. During ecdysis the 
head bowed downward and under towards the pros- 
temum until the abdomen completely withdrew from 
the exuviae. The apex of the abdomen thrashed from 
side to side to shake off the peh. The unexpanded 
wings were held out and away from the thorax and 
bent caudad parallel to, but not touching, the sides of 
the thorax. By 1 0:20 PM the wings reached the fourth 
segment. The expanded bases were cream white while 
the unexpanded apices remained smoke black. 

In a new species of Dactylocerca from Alamos, 
Mexico, an adult emerged at 1 0:40 PM and by 1 1 : 1 5 
PM its wings had already assumed their definitive 
shape. They were pure white and no veins or pattern 
were evident. The exuviae was not yet consumed. 
Surprisingly, while still teneral, the adult was able to 
resume its silk spinning. By 8:00 AM the next day 
the male was semi-hardened but the wings were sfill 
pale. By this time the exuviae was almost entirely 
ingested. At 8:00 AM, the third day, the male was 
still in the same place, fully pigmented, and the exu- 
viae had been completely consumed. 



Parthenogenesis 

Parthenogenesis probably occurs sporadically 
throughout the order but only a few cases, involving 



Mediterranean and equatorial African species, have 
been studied. One of these, Parthenembia reclusa 
Ross ( 1961 ), is widespread in western Africa. More 
recently, I have decided that a number of undescribed 
bisexual species from Angola and southeastern Zaire 
must be assigned to Parthenembia, therefore, this 
generic name is inappropriate. Scelembia virgo Ross 
( 1 960) from Angola and Zaire also is parthenogenet- 
ic. Other species of this potentially large genus are 
bisexual. Caryiology of the above two species was 
investigated by Renzo Stefani of Sardinia (1961). 

Parthenogenesis of Haploembia solieri (Rambur), 
indigenous to the Mediterranean region, was inten- 
sively researched by Stefani and reported in a series 
of papers cited in his bibliography. Rosaima Gior- 
dano of the University of Vermont, who is fluent in 
Italian, kindly provided the basis for the following 
summary of Stefani's conclusions on the bisexual and 
parthenogenetic forms of//, solieri: 

Stefani noted that bisexual H. solieri has only 20 
chromosomes whereas the parthenogenetic form has 
22. He assumed that the sex chromosome had been 
replicated. 

Haploembia solieri, like many other insects, suf- 
fers infection by the gregarine sporozoan parasite, 
Diplocystis clerci. Such infections aren't well toler- 
ated in bisexual populations of//, solieri for, although 
both sexes are debilitated, the influence on males is 
critical for they become ineffective mates due to dam- 
aged sperm and lowered vitality. In contrast, a mi- 
nority of females in a population which can repro- 
duce parthenogenetically, tolerate infection and, even 
though they produce fewer eggs, they can at least re- 
produce. As a result, parthenogenetic females sup- 
plant sexual populations, as has happened on Sardin- 
ia and Corsica. 

During the 1960s when Stefani conducted his re- 
search, small islands, such as Elba, had both sexual 
and parthenogenetic individuals. He predicted that 
with time only parthenogenetic populations would 
exist on these islands. Perhaps by now this transfor- 
mation has occurred. 

Because the parthenogenetic form tolerates infec- 
tion, and thereby permits D. clerci to mature and com- 
plete its life cycle, //. solieri can serve as a vector for 
the parasite. 

Another of Stefani's interesting observations was 
that virgin females of the bisexual form of//, solieri, 
unlike those of two other species of Haploembia he 



n 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES. NO. 149 



checked, have the habit of eating their unfertihzed 
eggs. Very few eggs escape this cannibahsm. 

Strangely, 22-chromosome females (parthenoge- 
netic), unlike 20-chromosome females (sexual), do 
not eat their own unfertilized eggs. Sometimes para- 
sites alter behavior of their hosts. However, infected 
parthenogenetic females aren't thus influenced. 
Stefani was inclined to conclude that this results from 
a duplication of the sex chromosomes, or perhaps a 
combination of both — an interesting question to in- 
vestigate. 

Obviously, Stefani wished to determine how par- 
thenogenetic females could have arisen from sexual 
females. As an experiment he deliberately infected a 
sexual female (20 chromosomes) with Diplocystis. 
It managed to lay an egg in 1 959, but this didn't hatch 
until 1961 ! The egg produced an adult female that 
began to oviposit without mating. Much to his sur- 
prise her progeny had a chromosome number of 22. 
In effect, he had artificially produced parthenogene- 
sis. He must have been amazed! [End of Giordano's 
abstract, somewhat reworded by me.] 

Because every individual is reproductive, the par- 
thenogenetic form of Haploembia solieri readily be- 
comes a "weed"; one rapidly spread and established 
in new lands by human commerce. From California 
where it was perhaps introduced in early Spanish 
"Hides and Tallow" commerce, including dumping 
ashore of sailing ship ballast. H. solieri is now a very 
widespread species in warmer habitats of southwest- 
em United States and northwestern Mexico. I have 
also found the parthenogenetic form common in many 
Mediterranean mainland localities. Bisexual H. so- 
lieri, and related species, or races, also occur, but these 
apparently ha\ e not yet become parthenogenetic as a 
result of Diplocystis clerci infection. 

Much to my surprise, a bisexual population of H. 
solieri was found in a garden (isolated from natural 
environments) south of San Francisco, California. I 
assumed that it resulted from a recent introduction in 
nursery stock, perhaps from Spain. I attempted to 
mate males with the ver>' common parthenogenetic 
females well established in nearby hills. The culmre 
produced only parthenogenetic broods. I didn't ob- 
ser\'e copulation but, even if this had occurred, it is 
likely that the parthenogenetic females were already 
"self-fertilized." This crossing attempt was made in 
1976. It is possible that the sexual population has since 
been eliminated by infection with D. clerci. 



Diet 

In nature embiids primarily eat weathered outer 
bark of trees and decomposing leaf litter. They may 
also eat mosses and lichens growing on bark, rocks, 
termite mounds and soil surfaces. Undoubtedly, many 
old substrates are coated or pemieated with live mi- 
croorganisms, such as algae, which are also nutritious. 
There is no evidence that digestion is dependent on 
symbiotic intestinal organisms. 

It is likely that the diet of embiids is primordial 
and that during the entire evolutionary history of the 
order there has always been a certainty of food wher- 
ever the insects choose to live on the basis of other 
environmental factors, such as availability of crevice 
retreats. Trees with exfoliating bark flakes, or verti- 
cal crevices, are most likely to have embiid colonies 
on them even though the nutritive value of the outer 
bark of other tree species in the environment might 
be the same. 

Embiids seem to have no host plant preferences, 
but one may expect that freshly-fallen leaf litter of 
plants with strong antiherbivore chemicals, as in As- 
clepias. Euphorbia, and Eucalyptus, will be avoid- 
ed. However, such litter can be assimilated if 
decomposing. For example, species of Australembi- 
idae feed almost exclusively on a diet of layered, ag- 
ing Eucalyptus leaf litter which also serves as the 
habitat (Fig. 23). In Antipaluria intermedia (T)a\\?,) 
of Venezuela, the dry season may be spent in leaf 
litter and the wet season in sheet-like colonies on the 
bark of adjacent trees. 

A highly neotenic new genus and species from 
the desert steppes of western Afghanistan extends 
foraging galleries upward from subterranean retreats 
into Anemesia shrubs to reach live foliage. In this 
case, the small, chemically-protected, aromatic leaves 
are transported into deep subterranean galleries as a 
food supply during periods when the surface envi- 
ronment is intolerably hot and dry. It can be assumed 
that similar provisioning occurs in other species of 
embiids inhabiting arid environments. 

The universal acceptance of any non-toxic dead 
leaves as food is demonstrated by my success in cul- 
turing hundreds of usually-unrelated species from all 
regions of the world on the same diet — dead Califor- 
nia live oak leaves and a supplement of fresh Ro- 
maine lettuce. Although not an essential element in 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 2 



19 



the diet, frequent replenishment of lettuce to the sur- 
face of a laboratory culture increases canying capac- 
ity in a limited culture space while also moisturizing 
the diet. 

Adult females simply continue the diet of nymphs 
of both sexes. Adult males, however — at least of all 
non-neotenic species — never feed. This conclusion 
is based on examination of thousands of adult male 
specimens of hundreds of species during KOH mac- 
eration while making microscope slides. These ob- 
servations show that the gut of non-neotenic males is 
invariably empty except during the short, teneral pe- 
riod when it only contains fragments of its own pen- 
uhimate nymphal exuviae which it ingested shortly 
after the final ecdysis. This peh gradually moves 
caudad in the otherwise empty gut and eventually is 
voided. Earlier workers, discovering such fragments 
in the gut, erroneously concluded that embiid males 
are predaceous. It is likely that all embiids, as do 
many other insects with chewing mouthparts, invari- 
ably ingest their exuviae after each moult. Some phys- 
iologists believe that this insures a beneficial recychng 
of sugars and nitrogen for chiton is a nitrogen-con- 
taining polysaccharide. 

Apparently, males of almost all species cease nor- 
mal feeding during the penultimate instar and com- 
pletely empty the gut prior to the final moult. This, 
together with a reduced, or arrested, accumulation of 
fat, results in a lighter, more vagile organism, but one 
with shorter life expectancy 

Neotenic, apterous, adult males of certain species 
may, or may not. have the intestinal tract filled with 
food in the usual stages of assimilation. Ingested 
exuviae would be less discernible in such accumula- 
tions. Males which continue to eat as adults are found 
in several new genera in central Africa and in a race 
of Metoligotoma reducta Davis of Queensland. Such 
males, having a more pronounced intraspecific de- 
gree of neoteny, have mandibles similar or identical 
to those of nymphs and females and thus they are 
suited for chewing food and apparently not for grip- 
ping a female's head during copulation. 

Perhaps due to genetically-fixed, behavioral traits, 
there is usually a correlation of taxon and specific 
habitat and its particular food resources. For exam- 
ple, in any suitable Queensland locality one encoun- 
ters at least one species of each of the three families 
occurring in Australia. Notoligotomids will be found 
on rock, ledge, and bark surfaces without gallery ex- 




FlGURE 23. Decomposing leaf litter o'i Eucalyptus is the 
typical habitat and food of the numerous species and races 
of Australia's family Australembiidae. Such galleries ap- 
parently don't extend deeply into soil beneath the litter. 
Metoligotoma ingens Davis. Black Mtn., Canberra, Aus- 
tralia. 



tensions into soil; australembiids will be encountered 
between layers of dead leaves (usually Eucalyptus) 
and, likewise, never extend galleries into soil; and 
oligotomids, having a more widespread Australian 
distribution, will be the only species dependent on 
soil retreats. However, in the laboratory, species of 
all three families thrive under identical cultural con- 
ditions and eat the same food. Comparable correla- 
tions of taxa and habitat may exist in any environment 
inhabited by embiids. Therefore, an experienced col- 
lector routinely examines each characteristic micro- 
habitat as a means of securing cultures of all species 
in the region. 

It is probable that embiids seldom face food or 
habitat limitations and this accounts in part for the 
absence of striking biological and anatomical diver- 
sity within the order. It is apparent that embiids are 
never able to fully exploit local environments, or to 
spread out into all suitable habitats. Typically, oc- 
cuirence in any environment is spotty and many an 
apparently satisfactory habitat, or even a major re- 
gion, seemingly lacks representation of the order. 

Such apparent absence may be attributed to haz- 
ardous dispersal. Flight, because it is limited to males, 
cannot extend a species' range and the almost com- 
pletely gallery-dependent, apterous females and 
nymphs can walk only short distances outside of the 
parent colony before encountering a predator — most 



20 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



likely an ant. Unlike nuptial tennites, ants, mayflies, 
etc., embiids are unable to overwhelm the predation- 
potential of an environment by concurrent bursts of 
thousands of dispersing nuptual individuals in a lim- 
ited time and space. 

Another factor is the lack of a need to disperse 
because of exhaustion of the food supply, or crowd- 
ing. This is especially apparent in tropical evergreen 
forests. For example, because food in the form of 
weathered bark and surface gi-owths would be restored 
very soon after consumption, it is possible that suc- 
cessive generations of one or more species of embi- 
ids could remain indefinitely on a single, large tree 
trunk. The activities of parasitoids. diseases, and other 
natural hazards would also tend to limit embiid pop- 
ulations. 

Because of reduction in dispersal incentives and 
limited vagility, embiids promise to serve as excel- 
lent indicators of zoogeographic regions and conti- 
nental drifit. 

Movement 

Embiids are especially adapted for movement in 
narrow galleries. Shortness of legs, probably a plesi- 
omorphic feature retained from a Paleozoic arche- 
type, is especially important. Such ancestors probably 
depended on bark, rock, plant crevices, or layered 
leaf litter, as reftjges from predators. Because such 
retreats usually are edible, embiids didn't have to 
venture far afoot, or in flight, to reach food. Howev- 
er, even short forays to extend "grazing" were made 
safer by evolution of an ability to produce silk cover- 
ways. 

Although wings were once possessed by adults 
of both sexes, rapid flight probably never was impor- 
tant as a means of avoiding predators. However, aerial 
dispersal of alate females of ancestral species must 
have fostered spread of taxa over a long period of 
geological time, thus extending the order's range 
throughout warm portions of Panagea. Later, how- 
ever, except for sporadic mo\ement of gallery sub- 
strate objects by storms, bird plumage, or in 
commerce, dispersal is severly limited by universal 
female apterism. Therefore, walking by females is 
the only way an embiid can move to a new location if 
the one presently occupied becomes intolerably wet, 
dry, exposed, or "overgrazed." 

Ordinarily, nymphs remain within their galleries 
unless they are torn open by predators, such as birds, 



mammals, or army ants. Adults of both sexes, how- 
ever, may concurrently leave their galleries soon af- 
ter the first rains following a long diy season, or during 
what appears to be primordial, prenuptual excitement 
comparable to that causing simultaneous nest-exo- 
dus of nuptual termites and ants. However, because 
embiid colonies are less populous and scattered, nup- 
tual embiids never create noticeable swanns. Dis- 
persal of individual embiids are therefore probably 
more vulnerable to predators. Consequently, some 
diurnal species of embiids have evolved a degree of 
protection by mimicking appearance and movement 
of unpalatable or dangerous models, such as ponerine 
ants, paederine Staphylinidae, or lycid orpyrochroid 
beetles. 

Forward walking, either within or outside of the 
galleries, is steady, slightly sinous, with all legs in- 
volved. Stimulation of the cerci triggers rapid for- 
ward bursts of speed. However, the most important 
defensive movement is rapidly backward, powered 
by the enlarged tibial flexor (depressor) muscles that 
almost fill the large hind femora. A firm tread, espe- 
cially on a silk surface, is insured by numerous, stout 
setae on the plantar surface of the hind basitarsi. 
Backward movement has had a profound influence 
on wing specializations, as well as being the primary 
cause of neotenic apterism and subapterism. 

During casual forward movement outside of gal- 
leries, adult females often walk with their genital seg- 
ments arched upwards as though to "welcome" 
insertion of a male's terminalia. Adult males of many 
species often walk with the abdominal apex bent for- 
ward on the dorsum of the abdomen. In alate species 
this apex may even press tightly against a correspond- 
ing forward fold of wing apices and must be an addi- 
tional means of reducing the adverse barb effect of 
both wings and tenninalia during reverse movement 
within galleries. Such forward bending of the ab- 
dominal apex also occurs in apterous males of some 
species. 

Males outside of galleries are very lively and alert. 
Often they stand high on the forelegs, thus elevating 
the usually large-eyed head, and the prothorax. A 
male's head is very mobile and capable of turning at 
least 45° from the longimdinal axis. Especially while 
resting on walls beneath artificial lights, males often 
twist the head in a mantid-like fashion as they follow 
the movements of an observer. Just before flight, a 
male's body may tremble, antennae vibrate and twirl, 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



21 



and the forebody bobs up and down (Fig. 1 1 ). Alate 
males thus retain much of the activity and sensitivity 
of free-living ancestors while the nymphs and adult 
females, perhaps because they are secluded in 
galleries, are less alert. 

Embiids are highly sensitive to vibration, as may 
be observed in laboratoiy cultures kept in jars. In 
these, adults often rest for long periods in upper gal- 
lery levels. In reaction to human approach, or vibra- 
tion, even as much as ten feet away, embiids suddenly, 
often in unison, back downward into the depths of 
the culture. Because of this, a person wishing to 
collect such individuals, must approach a culture jar 
slowly, gently open its lid, and trap the desired indi- 
vidual by blocking the gallery behind it. At times 
embiids feign death, even during handling, and then 
suddenly burst into activity. 

Habitats 

1. TROPICAL EVERGREEN FORESTS. 

Such forests appear to be the basic, or primordial en- 
vironment of the order. Most species occurring in 
wet forests are arboreal, or colonize sheltered sur- 
faces, such as undersides of ledges, logs or branches 
which remain relatively dry during frequent rains. 
Even on well-drained, vertical tree trunks, or road 
banks, the sheltered slant-side is prefen'ed (Fig. 24). 

In spite of the great number of species potentially 
present in a particular forest with its profusion of 
microhabitats and food, an experienced collector may, 
after hours, even days of concentrated search in a vir- 
gin forest, fail to find a single colony. This difficulty 
apparently results from the abundance of ant preda- 
tors, diffusion, and an inability of embiids to fully 
exploit or reach potential habitats. 

An entomologist, regardless of his specialty, is 
well aware that the best general insect collecting is 
found in recently disturbed forests and this is partic- 
ularly the case with embiids. One soon learns that 
embiid colonies are most frequently encountered on 
residual trees, stumps and logs left in forest lands 
partially cleared (but not burned) for plantations, such 
as those of cocoa and coffee. Also, clearance and 
trails provide easier collector access to potential sites. 

Natural and artificial habitats within tropical for- 
ests are listed as follows: 

(a) Surfaces of trunks, limbs and lianas. Some 
embiid species spin conspicuous galleries of clean 
silk fully exposed to view (Figs. 25, 26). As weath- 



ered outer bark is consumed, galleries are extended 
over fresh surfaces. Occasionally, an entire tree trunk 
is matted with the silk of apparently merged galleries 
of separate broods (Fig. 26). It is conceivable that 
regrowth of an edible substrate is delayed by such 
cover and that portions of a trunk will have to be aban- 
doned. While large colonies are active, the matting 
of silk constitutes a protective cover, which, as stated 
before, is almost as effective as a layer of bark for it 
not only protects new galleries, but also other organ- 
isms taking advantage of the cover. 

Often, for some distance around a large colony, 
sinall satellite colonies will be found on adjacent trees. 
Farther out, the forest may lack additional colonies 
until another concentration occurs. In some species 
gallery silk on bark is made inconspicuous by a cov- 
ering of pulverized fecal pellets and bark fragments 
deliberately placed on the surface by the embiids 
(Figs. 8, 9). Often minute plant life will grow on this 
and render the silk even more obscure while also en- 
hancing the protective cover. 

(b) Bark crevices and flakes. At times galleries 
of small species are not visible without removal of 
bark flakes. However, presence of a colony is often 
indicated by a slight extension of silk beyond the edge 
of a flake. 

(c) Roots of orchids and other epiphytes. Vines 
and their leaves appressed to tree trunks often shelter 
colonies. For example, the flat, circular leaves of 
Peperomia rooted to tree bark provide excellent cov- 
er for colonies of Saiissiirembia Davis in Costa Rica. 

(d) Undersides of elevated logs often criss- 
crossed in clearing and selectively-logged forests. 

Recently-felled trees provide access to colonies nor- 
mally out of reach on standing trees. Because of 
changed exposure, the colonies may shift to the un- 
dersides of levelled trunks and branches, thereby gain- 
ing greater shelter than that afforded on upright trees. 

(e) Fence and stockade posts. Especially those 
recently cut which still have loose bark, are frequent 
sites of colonies. 

(0 Surfaces of ledges and earthen banks. Es- 
pecially if somewhat under-slanted and fa\'orably ex- 
posed toward or away from the sun, these surfaces 
tend to serve much like tree bark in that they otfer 
many retreat crevices and a food supply in the fomi 
of surface growths. 

(g) Surfaces and crevices of structures. E\ en 

mossy, steel girders of bridges. 



22 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure 24. Typical rainforest habitat (semicleared). Several embiid species may be 
found on such a tmnk. Near Belem. Brazil. 



2. TROPICAL CLOUD FORESTS These oc- 
cur at various altitudes, usually starting at 3,000 
meters on wet, often windward slopes of tropical 
peaks and ranges. Embiids reach their highest known 
altitude in this zone — about 3,500 ineters in the Ec- 
uadorian Andes. In spite of cold nights and proxim- 
ity to snow, each day is warm in equatorial latitudes. 
Therefore, there is no need for the special, low tem- 
perature, physiological adaptations, required in tem- 
perate regions with prolonged cold periods. 

Males of many cloud forest species are slender- 
bodied, have disproportionately long legs and large 
wings (Fig. 12). All are difficult to culture and it is 
advisable to secure adequate series of adults in the 
field by persistent collecting rather than to depend 
on culturing. Late instar nymphs may survive long 
enough to mature, however. 

All habitats listed for lowland rain forests should 
be searched for embiids. However, colonies are not 
only less conspicuous, but also less accessible, be- 
cause of steep ten'ain. usually impenetrable vegeta- 
tion (especially bamboo), and dense co\'erings of 



mosses and epiphytes on most surfaces. Because of 
such disadvantages, the best collecting opportunities 
are in agricultural clearings and in edges of residual 
gallery forest. 

The most characteristic cloud forest einbiid hab- 
itat is moss festooning from tree liinbs and trail banks. 
Such inoss is often thoroughly bound together with 
embiid galleries (Fig. 27). 

Because of better drainage, shelter, and accessi- 
bility, road and trail banks are the best places to search 
for embiids in cloud forests. At times galleries on 
road banks can even be seen from a moving vehicle. 

3. SEASONALLY-DRY GRASSY WOOD- 
LAND. In the rainy season these forests are lush and 
green, but the long dry season and repeated fires tend 
to reduce epiphytic growth and thus suitability of most 
tree trunk surfaces as embiid habitats. However, this 
zone — especially the savanna woodlands of central 
Africa and the deciduous woodlands of India and 
southeast Asia (Fig. 28) — is home to some of the most 
interesting species of the order. It should be noted 
that evergreen groundwater forests frequently occur 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



23 




Figure 25. In rainforests, colonies often are conspicuous. 
It is surprising that they are so often neglected by ento- 
mologists. Clothoda longicauda Ross. Tingo Maria, Peru. 




Figure 26. Interconnected galleries of many broods of 
Machadoembia n.sp. almost cover buttresses of huge trees 
in Congo rainforests. Such concentrations occasionally are 
ripped open and plundered by vertebrate predators. 



within this type of woodland along rivers and on shad- 
ed slopes. The characteristic dry woodland habitats 
are: 

(a) Tree bark crevices and flakes. Several large, 
pale species are found in bark, but tend to develop in 
individual galleries, nymphs having dispersed soon 
after hatching. At certain seasons only tiny nymphs 
of potentially-large-bodied species will be found only 
after careful search. Species of African genera, Ber- 
landembia Davis and Dinembia Davis, are exam- 
ples. Occasionally, seemingly-empty, prior-season 
galleries will be found, but close exainination will 
reveal unattended live egg masses or tiny nymphs 
enduring the dry season, above the height of bush 
fires. 

(b) Dead branches attached to trees. These are 
especially impoilant habitats. In them, man-made 
splinter-cracks, beetle burrows, pithy twigs and loose 
bark serve as refuges from predators, temperature 
extremes, and dessication. 



(c) Dead branches and logs on ground. Espe- 
cially those recently cut or fallen should be rolled 
over and their undersides carefully examined. Gal- 
leries once connected with those in thus-exposed leaf 
litter and soil crevices should also be searched. 

(d) Leaf litter on forest floor. Embiid galleries 
from deep, protective, soil crevices may extend into 
such cover during the wet season, or cooler periods 
of the day or night. Species of the Australian family 
Australembiidae almost exclusively occur amongst 
matted leaves and apparently do not utilize soil re- 
treats (Fig. 23). 

(e) Crevices in surfaces of rocks, ledges and 
termite mounds. The latter offer almost rock-hard 
protective retreats, good drainage in the wet season, 
and the weathered surfaces are often rich in nutrients 
(Figs. 5, 6). 

(f.) Leaf clusters in understoiy of savanna wood- 
land {Figs. 2S\ 29 A, B). 



24 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




Figure 27. In cloud forests many unrelated species utilize 
well-drained, festooned moss both as habitat and food. New- 
genus and species (Oligotomidae). Gunong Batu 
Brinchong. Malaya. 



4. SEMI-ARID, OPEN GRASSLANDS When- 
ever present, all previously mentioned habitats should 
be searched in this zone. Forest habitats may be 
localized along river courses and around springs. 

(a) Stones. Most grassland species are best en- 
countered under stones (Figs. 30, 3 1 ) but it should be 
realized that the turned stone simply exposes a soil 
profile (Fig. 32) for it is likely that most grassland 
embiids are widely distributed in the sod and are not 
necessarily stone-cover-dependent. Very little sur- 
face activity will be seen in the dry season and exca- 
vations may be required to secure cultures (Fig. 33). 
However, there is a chance of encountering embiids 
in surface galleries at night or early in the morning. 
Their attraction to the surface could be increased by 
aitificially wetting soil around galleries expected to 
contain embiids. 

(b) Dead branches and limbs riddled with bee- 
tle borings. Some embiids utilize these retreats to 
survive the long dry season and fires. 

(c) Leaf litter and soil. At the bases of large trees 
and beneath clumped shrubs. 

(d) Soil crevices in open ground. In western 
Australia, immediately after rains, species of 
Aposthonia extend galleries in soil crevices upward 








FlGLRE 28. The sa\ anna (Brachvstegia) woodlands of the Congo-Zambezi di\ ide have 
the greatest diversit>' of higher embiid taxa. "Nests" of female Enveja (Fig. 29A) were 
present on low shrubs in this scene. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



25 




Figure 29A. Silk webs in a leaf cluster. Such leaves may 
be dead or alive on low understory vegetation. This habitat 
is chiefly used by species of Enveja Navas occurring in cer- 
tain portions of Africa's Brachystegia woodlands. By oc- 
cupation of leaf clusters, females and their early brood es- 
cape excessive soil moisture in the rainy season. With rain- 
fall decline, the embiids move down into leaf litter and, fi- 
nally, into soil cracks to escape dry conditions and fires. 




Figure 29B. OpcncJ Lm^ju nc^i revealing female guard- 
ing eggs. Unlike those of most embiids, the eggs are loosely 
covered with fibrous debris — perhaps her fragmented fe- 
cal pellets. 



to reach leaf litter food. Fragments may be carried 
down into subterranean galleries to serve as food pro- 
visions. With a return to complete aridity, temporary 
surface galleries may soon weather away and thus no 
evidence of embiid occurrence may remain. An ex- 
ceptional species in western Afghanistan extends gal- 
leries upward into low shrubs {Artemesia) to collect 
leaf fragments. 

5. DESERT AREAS. An embiid fauna is consid- 
erably reduced by extreme aridity and may be con- 
fined to oases, drainage lines and foggy coastal 
deserts. A lack of significant precipitation almost 
completely eliminates occupation of all above-ground 
habitats. However, palm trees, such as the date palm, 
provide safe retreat in leaf bases of the trunk which 
tend to collect abundant embiid food in the form of 
leaf debris. In the Nile Delta two species of the or- 
der, Embia savignyi Westood, and Haploembia solieri 
(Rambur), occur in nest material deep in the under- 
ground burrows of rats of the genus Arvicanthiis 
(Hoogstraal, pers. com.) (Figs. 34-36). 

6. HUMAN HABITATS. Cities may have large 
embiid populations on the bark of shade trees. Most 
often these are introduced species of the India-cen- 
tered genus Oligotoma: O. saiindersii (Westwood), 
O. humbertiana (Saussure), and O. nigra (Hagen). 
Native species may also be found, especially if the 
highly disturbed areas are near natural environments. 
In Trinidad extensive webs of a native species, 
Antipalitiia wichi (Saussure), are conspicuous on 
trees along streets and in parks. 

Occasionally, embiids eat stored products. For 
example, species have been reported in sugar refin- 
eries in Australia. 

In Peith, Australia, Notoligotoma hardyi (Frieder- 
ichs) inhabits old wood fences in residential areas uti- 
lizing crevices as retreats and weathered wood 
surfaces as food. In southern California galleries of 
Oligotoma nigra Hagen occasionally extend fi-om the 
ground up the sides of residential foundations, but 
embiids do no damage to structures. 

In vineyards embiids may extend their webs into 
bunches of grapes near the ground, but probably feed 
only on dead material and detritus accumulated in 
interstices, not the grapes. In Senegal. Oligotoma 
saimdersii (Westwood) may produce colonies in piles 
of harvested peanuts accumulated prior to export. 



26 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 




■^^^ 



Figure 30. An excellent slony. seasonally-dry embnd habitat m southern India. (David 
Cavagnaro, assistant during Indo-Australian expedition). 





Figure 3 1 . Ihu dry slope m ccnual .Algeria, loo sloiiy loi 
plowing, is the type locality of Embia silvestrii Davis 
(Embiidae). 



Figure 32. A removed stone exposes a sod profile with 
galleries of Haploewbia solieri (Rambur) (Oligotomidae) 
in Califomian seasonally-dry grasslands. From cool moist 
depths, embiids move upward to feed on dead plant litter, 
perhaps at night or other cool periods. 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



27 




Figure 33. In arid Baja California. Mexico. Bii/hocerca 
sini (Chamberiin) (Anisembiidae), feed at night and dur- 
ing the rainy season, on leaf htter accumulated between 
desert stones. Dense silk galleries extend downward in the 
silty soil. 



Figure 34. Numerous galleries of Notoligotoma n. sp. 
(Notoligotomidae) are found, but not exclusively, on 
undersurfaces of exfoliating slabs on "granite islands" in 
southwestern Australia. Wind-blown plant debris caught 
beneath such slabs is the basis of an ecosystem which in- 
cludes many embiid colonies. 




Figure 35. Deserts near sea coasts may have a special 
embiid fauna supported by fog precipitation. For example, 
in the almost rainless coastal fog zones of Peni and south- 
western Angola, embiids may extensively colonize fog-sup- 
ported lichens growing on rock and ground surfaces. Re- 
treats beneath rocks and into soil and rock crevices pro- 
vide escape from the excessive dryness of certain seasons 
and/or hot periods of the day. Illustrated is the habitat of 
Chelicerca n. sp. (Anisembiidae) in the Peruvian Desert 
north of Callao. 




Figure 36. Galleries of some species occur on the sur- 
faces of roots of perennial shrubs as well as of trees. An 
interesting new genus and species was thus found in Africa s 
Namib Desert following the roots of Wehvitschia plants. 
However, the species probably isn't restricted to a host. 



28 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Geographic range 

Embiids are warm-climate-adapted insects whose 
natural occurrence is almost universal in all suitable 
environments on any continent, or continental island, 
which has a tropical or wami-temperate climate. The 
order's distribution roughly coincides, for example, 
with that of phasmids and termites but the latter ex- 
tend farther into colder latitudes, perhaps because of 
deeper penetration into habitats, such as soil, logs, 
and buildings. But, as with termites, embiid occur- 
rence in temperate regions is reduced. In western 
North America the most northerly record of an en- 
demic species, DacU-locerca rubra (Ross), is in cen- 
tral Utah, lat. 39°N. Non-endemic Haploembia solieri 
(Rambur) occurs farther north into eastern Washing- 
ton. In eastern North America, represented by Dira- 
dius vandykei (Ross), the order reaches about 35° N 
in coastal Virginia. 

In South America the southern limit has not been 
determined. The most southern occurrence known 
to me is Pararhagadocbir trachelia (Navas) in a 
desert habitat west of Mendoza, Argentina, at about 
35° S. West of the Andes, the southern range appears 
to be limited by the extreme aridity of the Atacama 
Desert of northern Chile, rather than by latitudinal 
cold. An undescribed species of Chellcerca Ross 
probably extends well southward in the coastal fog 
belt (lomas) of northern Chile, but, to date, 1 have 
found it only as far south as Mollendo, Peru. The 
northern Chilean lomas aren't very accessible to a 
collector. There is absolutely no evidence of a South- 
em Hemisphere, or Patagonian, origin of any spe- 
cies. 

In the Old World, endemic species of the order 
occur in warni portions of all continents and conti- 
nental islands even including Tasmania [Mero- 
Ugotoma tasmanica Davis and Aposthonia gurneyi 
(Froggatt)]. New Zealand and Madagascar, which 
apparently could not be reached by natural spread of 
any species, evidently do not have endemic repre- 
sentation of the order The northern extremity of range 
in western Eurasia is in the Crimea, about 45°N, with 
Haploembia solieri (Rambur). In the Middle East, 
Parembia persica (McLachlan) ranges into northern 
Iran, Afghanistan, and Turkestan (about 38°N). In 
eastern Asia, Aposthonia japonica (Okajima) reach- 
es about 32°N in southern Japan, but this may be a 
species introduced from a south- Asian locality^ It is 
assumed that the order ranges to comparable latitudes 
along the coast of mainland China. 



Oceanic islands of the Pacific appear to have no 
endemic species, those present having been intro- 
duced by movements of ancient and modem man. 
Aposthonia Oceania (Ross) is widespread in Oceania 
as far south as Easter Island 27°S and New Caledonia 
(20°S ), at first by dispersal of Polynesians. 

The fossil record indicates occurrences far to the 
north of present ranges, but this may be due to drift 
of land surfaces northward from warmer zones sub- 
sequent to fossilization. In the case of the Florissant 
fossil, Lithembia jlorissantensis (Cockerell), the fos- 
sil beds probably drifted northwestward and were lift- 
ed an additional 5,000 feet in altitude by elevation of 
the Rocky Mountain Range. I speculated that Baltic 
Amber fossil. Electroembia antiqua (Pictet), lived 
when the Baltic land surface was in a wami, dry Med- 
iterranean latitude (Ross, 1956). 

Ecological range 

Within the extremes of geographic range, embiids 
are confined to regions lacking prolonged cold peri- 
ods; the more equatorial the location, the greater the 
potential altitudinal range. Aridity can also be a lim- 
iting factor, but some practically rainless deserts, such 
as the Pemvian and Namib, have one or more embiid 
species in habitats regularly dampened by sea fog. 

The primordial habitat of the order appears to be 
tropical rain forests. Those of the Amazon Basin re- 
tain the order's most "primitive" species. From such 
forest centers, embiids appear to have radiated and 
become adapted to many other types of environments. 
In this movement the trend has been toward an eva- 
sion of adverse ecological conditions rather than phys- 
iological adaptation to them. Thus, in regions 
experiencing prolonged dry seasons, embiids escape 
heat, dessication and fires by retreating downward in 
silken galleries following cracks, which provide ac- 
cess to cooler, more moist soil depths. Others may 
find refuge in burrows produced by beetle lar\'ae in 
dead limbs of trees. Under such circumstances the 
activities of the embiids are confined to the rainy sea- 
son or to periods of the day or night when the surface 
temperature and moisture are favorable. 

In equatorial regions embiids spreading from op- 
timum lowland zones have adapted to higher eleva- 
tions. The highest known altitudinal record of the 
order is that of a new genus and species found in cloud 
forests and paramos above Cuenca, Ecuador at about 
3,500 meters elevation. The order is present in most 



ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



29 



of the cloud forests I have visited, e.g., Andean, cen- 
tral African, and Malaysian. In such high zones the 
low temperature extreme may not be much more se- 
vere than that of many warm-temperate regions at sea 
level. The zones are also subject to almost daily in- 
tense equatorial, solar heat at all seasons. A few 
endurable hours of cold occur one night at a time rath- 
er than for weeks and months, as in cold latitudes. 

Low temperature, however, appears to be the most 
critical ecological factor limiting the order's occur- 
rence. Perhaps this is due to an absence of a truly 
cold resisting or hibernating life stage. No species is 
known to utilize the egg stage as a means of surviv- 
ing cold periods, and, of course, there is no domiant 
pupal stage. Three species endure rather severe win- 
ter cold: Dactylocerca rubra (Ross) in central Utah, 
Anisembia texana (Mel. ) in northern Texas and south- 
em Oklahoma, and Haploembia solieri (Rambur) in 
Crimea and, by introduction into eastern Washing- 
ton. Haploembia solieri overwinters as half-devel- 
oped nymphs which, during cold periods, are 
protected in dense cocoon-like silk enclosures. Such 
a protection probably is characteristic of all cold-en- 
during species of the order. These "cocoons" proba- 
bly are a means by which predation is reduced during 
the embiid's torpid state for, in cold environments, 
potential predators, such as carabid beetles, are like- 
ly to be adapted for movement during cold periods 
and therefore can easily catch unprotected prey. 

Natural dispersal 

The major distribution of the order must have oc- 
curred on portions of Pangaea during the long period 
when both sexes possessed wings. After females be- 
came universally apterous, natural dispersal outside 
of galleries became slow and hazardous, especially 
after ants evolved to become significant predators. It 
is possible, however, that the natural spread of small 
species, such as teratembiids, could have occurred as 
substrate objects, such as twigs and branches, were 
moved by wind, and rafting. On or in such objects 
survival and transport of embiids would have been 
insured by enclosure in securely attached silk galler- 
ies. When continents, such as South America and 
Africa, were still separated by a narrow sea, aerial 
movement and rafting would have been more likely 
considering the long time periods available. 

Another, yet rare means of dispersal could be in 
bird plumage, as is likely to have happened in the 



case of Chelicerca galapagoensis Ross. Colonies of 
this anisembiid frequently web the general environ- 
ment and nests of birds in higher elevations in larger 
islands of the Galapagos. It would seem to be a sim- 
ple matter for embiids to extend their galleries into 
the feathers of birds resting on a nest and, although 
many of these would be groomed out of the plumage, 
a certain percentage over the ages, could be carried 
from place to place and thus experience an extension 
of range. Chelicerca galapagoensis is related to a 
Peruvian and Ecuadorian loma (fog zone) Chelicer- 
ca and it is likely that a limited gene pool of a future 
new species was carried to the islands from the main- 
land as "passengers" in the plumage of birds (Ross, 
1966). 

Dispersal by man 

Because of the use of crevice retreats and the ad- 
herence of silk galleries to logs, plant materials, ship's 
ballast and cargo, the range of some species has been, 
and will continue to be, extended when such objects 
are moved either by natural or artificial means. In 
warmer regions, especially where man has engaged 
in commerce for thousands of years, it is probable 
that species which appear to be endemic, were actu- 
ally unintentionally introduced by humans. This is 
particularly likely in southern Asia and the Mediter- 
ranean region. It is to be expected that the greater 
speed and volume of modem commerce will acceler- 
ate artificial spread of additional species. 

Embiidina are able, at least temporarily, to be- 
come established in greenhouses located in temper- 
ate regions. One of the earliest known species, 
"Embia'"' michaeli McLachlan, 1877, was collected 
in an orchid house in England and is known only from 
its incomplete type specimen. The origin of the or- 
chid appears to have been northeast India, or Burma. 

The establishment of a South American teratem- 
biid, apparently Diradiiis intricatits (Davis), in an 
orchid greenhouse near Wageningen, Netherlands was 
reported to me by R. H. Cobben. It was first noted 
about 1970 and, in spite of several intensi\'e insecti- 
cidal treatments, the infestation persisted many years 
until the greenhouse was removed for other reasons. 
It is unlikely, however, that such an introduction would 
have economic significance. Undoubtedly the insects 
were unintentionally introduced on live plants, per- 
haps orchids, from Suriname. 



30 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



For many years embiids were frequently inter- 
cepted by U. S. plant quarantine inspectors in orchid 
plants from various Neotropical sources — particularly 
Colombia. It is unlikely that any of these would have 
survived very long, even if they slipped past inspec- 
tion. 

Some embiid species are more successful than 
others in becoming established abroad and it is sig- 
nificant that most of these belong to the family Olig- 
otomidae. Certain species of this family often have a 
high survival potential and reproductive vigor as in- 
dicated by the fact that they are the easiest to propa- 
gate in laboratory cultures. They also tend to produce 
overlapping generations throughout the year. 

In order of importance, species whose range is 
steadily being extended by man are listed, as follows 
(Species 1-9 are in Oligotomidae): 

1 . Oligotoma saimdersii (Westwood). 

Endemic to northem India. Now likely to be found 
in any warm, moist tropical or warm-temperate re- 
gion. This is the species most often collected in set- 
tled areas by the non-specialist. Males frequently are 
attracted to lights. 

2. Oligotoma humbertiana (Saussure). 

Endemic to India. Now very common in Indone- 
sia, Philippines. China, western Mexico (probably in- 
troduced from the Philippines in galleon trade to 
Acapulco). Unfortunately, this species is becoming 
distributed in natural areas of western Mexico, espe- 
cially in the states of Sinaloa and Nayarit. Males fre- 
quently appear at lights. 

3. Oligotoma nigra Hagen. 

Probably endemic to northem India, O. nigra has 
spread westward and northwestward in ancient cara- 
van traffic and now is well established in Arabia, the 
Middle East and the valley of the Nile. It was appar- 
ently introduced into southern United States in date 
palm cuttings (Ross, 1 957) and now extends into east- 
em Texas and northwestem Mexico. It has also been 
collected m inland NSW Australia. It is now occu- 
pying natural habitats. Males frequent at lights. 

4. Oligotoma greeniana Enderlein. 

Probably endemic to India. Now established in 
Malaya, Taiwan, Mindanao, Hong Kong, Macao 
(probably from Goa). and China's southwestern in- 
terior. 



5. Aposthonia borneensis (Hagen). 

Probably endemic to continental southeastern 
Asia, or Indonesia. Now also occurs in southern 
China, Vietnam, Philippines, New Guinea and Indo- 
nesia. 

6. Aposthonia Oceania (Ross). 

Possibly endemic to continental southeastern Asia, 
or Indonesia. Apparently was spread by early Polyne- 
sians to various portions of Micronesia and Polyne- 
sia, including Easter Island and New Caledonia. In 
at least one case, Aposthonia micronesiae (Ross), 
from Mariana Island, speciation may have since oc- 
curred but there is a possibility that this is a species 
introduced from a source-region distinct from that of 
A. Oceania. 

7. Aposthonia ceylonica (Enderlein). 

Probably endemic to southern India and Sri Lan- 
ka. Now spread to Mauritius, Madagascar, Malaya, 
Thailand and Laos. 

8. Aposthonia minuscula (Enderlein). 

Probably endemic to India. Now spread to East 
African coast and Madagascar. 

9. Haploembia solieri (Rambur). 

Endemic to some undetermined Mediterranean 
area. Now common almost throughout Mediterranean 
shores and Black Sea region. This wide range per- 
haps resulted from ancient land and sea commerce. 
A parthenogenetic form occurs in scattered Mediter- 
ranean regions, especially on certain islands, and has 
since been spread to southem Spain, northwestem 
Aftica, Asia Minor, Afghanistan, the Canary Islands 
and westem United States — especially California. 

1 0. Embia savignyi Westwood. 

Probably endemic to southem Sudan and adja- 
cent regions. Now spread westward as far as north- 
em Nigeria and north into the Nile Delta. Israel, and 
Crete. Males fly to light. 

1 1 . Parembia persica (McLachlan). 

Probably endemic to northwestem India, or Pa- 
kistan. Now spread, perhaps at first by ancient com- 
merce, as far north as Russian Turkestan and as far 
west as Israel, Jordan and Arabia. The damaged type 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA. PART 2 



31 



of Parembia prodiicta Davis from Somalia appears 
to represent P. persica (Ross, 1981). 

12. Pararhagadochir trinitatis (Saussure). 

Probably endemic to northeastern South Ameri- 
ca and Trinidad. Now sporadically common on tree 
trunks around San Jose, Guatemala; northern Costa 
Rica; Panama; and probably elsewhere. 

13. Some South American species, such as 
Archembia batesi (Mc.L.) and Pararhagadochir hic- 
ingillata (Enderlein). appear to have extended their 
range in river commerce. 

Natural hazards 

PREDATORS: Outside of their galleries embi- 
ids are easy prey to predators, especially ants, as well 
as spiders, beetles, centipedes, and small vertebrates. 
Predation appears to be the prime factor limiting es- 
tablishment of colonies in new places. Out of their 
galleries it is difficult to think of creatures more vul- 
nerable to predation than embiids. They cannot run 
or fly very well, they lack biting or stinging defenses, 
repugnant secretions, body armor, n'ritating vestiture, 
or defense through high reproductive potential. There 
is indication, however, that some diurnally-dispers- 
ing species may reduce predation by mimicking the 
coloration and movement of repugnant or dangerous 
models. For example, in the Neotropical region many 
species, especially females, resemble notoriously ag- 
gressive, stinging ponerine ants. Most spectacular is 
the orange-and-black pattern, especially of nymphs 
and females, of several species of Dihybocercus 
Enderlein in south-central Africa. Combined with a 
similar body form, such embiids strongly resemble 
aposematic, highly initating staphylinid beetles of the 
genus Paederiis. 

The most significant defense, although not com- 
plete, is confinement in silk galleries. The impor- 
tance of silk gallery protection was clearly evident 
during a brief field experiment conducted by me in 
Singapore with colonies of Oligotoma saiindersii 
(Westw.). The colonies occurred on bark of shade 
trees on which swarmed large, vicious Oecophylla 
ants. These ants ran over the silk webs apparently 
unable to detect embiids underfoot. However, when 
I opened some webs, the exposed embiids were im- 
mediately detected and carried off by the ants. Un- 
der such circumstances the spread of the embiids 
would be possible only by gallery extension, or by 




Figure 37. Mirid-like appearance of a species of 
Plokiophilidae bug, Embiophila (Acladina) africana 
Carayon, I collected in Katanga, central Africa. A. Male. 
B. Brachypterous female. (After Carayon) 



movement outside of galleries during inactive peri- 
ods of predators. It is likely that ants are the most 
significant predators of embiids and that the major 
extension of embiid occuiTence on Earth took place 
before ants evolved and became abundant. Of course, 
during the pre-ant period, wide dispersal was aug- 
mented by presence of wings in both sexes. 

In addition to arthropod predation, colonies may 
be ripped open and plundered by birds, rodents, le- 
murs, monkeys, and other vertebrates. This is most 
evident in tropical forests where one occasionally 
encounters sheet-like mats of galleries torn open and 
devoid of occupants following discovery by an in- 
sectivore. 

As is to be expected in any ancient group of in- 
sects, embiids provide both habitat and food for par- 
asitoids and parasites, some exclusive to the order. 
Included are various species of tiny, brown, 
anthocorid-like Hemiptera of the little-know n family 
Plokiophilidae (Figs. 37, 38). All members of this 
tropical family live on silk webs; those of the sub- 
family Plokiophilinae appear to be confined to spi- 
der webs and the Embiophilinae to galleries of Em- 
biidina. Apparently, however, a substrate of silk and 
available food is more important than actual associa- 
tion with spider and embiid hosts for Carayon reared 
the bugs for several generations on a diet of mites. 
Carayon 's work ( 1974) is the best reference. 



32 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 





Figure 38. Plokiophilid bug nymph sucking fluids from 
the basitarsus of a dead embiid {Archembia n. sp., Tingo 
Maria, Peru). 



I have encountered the bugs sporadically in colo- 
nies of many species of embiids in both Old and New 
World tropics. The family appears to be absent in 
Australia. Often they are inadvertently collected as 
silk galleries and other habitat material is included in 
a field culture. As embiid cultures grow, embiophil- 
ines usually thrive, but apparently do not always kill 
their hosts. The bugs are most numerous in the vi- 
cinity of embiid egg masses and broods of young 
nymphs. Carayon(1974) reported that they regular- 
ly suck fluids from eggs and young nymphs. Tyro- 
glyphid mites are always present in embiid cultures 
and must supplement the bug's diet. I have also ob- 
ser\'ed them sucking shri\'eled. recently dead, embiids. 
Unless salivary toxins are introduced in their feed- 
ing, the bugs should not adversely affect the embiids 
unless several individuals simultaneously feed on a 
single egg or a small nymph. 

The following arthropods are also encountered 
within embiid galleries: 

Acarina. Predator}' mites frequently infest cul- 
tures and can seriously reduce their vitality. Larval 
mites of the family Trombiculidae attached to embiids 



have been encountered. Scavenger mites are also 
abundant in most colonies. 

Rhaphidioptera. The sinuous, predaceous lar- 
vae of snake flies frequently are encountered in un- 
der-stone galleries of Haploembia solieri (Rambur) 
in California. 

Coleoptera. On one occasion {Embia sabulosa 
group, S.W. Africa) I encountered numerous larvae 
of a species of Lampyridae which had consumed all 
of the embiids in an under-stone colony. I found a 
beetle of the family Monommidae in galleries of 
Clothoda longicauda Ross at Tingo Maria, Peru. 
Myers (1928) reported numerous monommids, 
Hyporrhagus marginatus (Fab.), within extensive, 
mat-like webs ofMesembia hospes (Myers) in Cuba. 

Diptera. Larva of Leptopteromyia Williston 
(Leptogastridae) have been reared on several occa- 
sions in embiid cultures from widespread Neotropi- 
cal localities. In the labyrinth of a culture it is difficult 
to determine the exact relationship of these slender 
fly larvae to embiids. However, frequency of emer- 
gence of adults from my cultures suggests that their 
larvae regularly utilize cover, and embiids as food, in 
field colonies. Carrera (1947) reported a puparium 
of the genus in embiid galleries in the botanical gar- 
den in Rio de Janeiro. 

ECTOR\R.\SITOIDS: The most interesting ec- 
tophagic parasitoids are larvae of wasps of the fami- 
ly Sclerogibbidae (Chrysidoidea). The only recorded 
hosts of this family are Embiidina. Male sclerogibbids 
are black, slender, alate, 1^ mm in length. Females 
(Fig. 39) are always apterous and move with great 
swiftness within the galleries, however, like their 
hosts, they spend much time resting motionless in 
upper levels of the galleries spun in culture jars. 

First instar larvae of sclerogibbids are legless, but 
otherwise resemble meloid triungulins. They are scle- 
rotic, well-segmented, and attach themselves to mem- 
branes in various parts of the host's body and, soon 
after beginning to feed, become maggot-like. As a 
rule, only one lar\'a is attached to a host (Fig. 40), but 
as many as four larvae of a large species have been 
found attached to a single embiid (Fig. 41 ). Larvae 
of certain small species of sclerogibbids may be more 
numerous on a single host; a dozen or more may be 
attached in a neat row on each side of the venter of 
the abdomen. 

When a sclerogibbid larva completes its feeding, 
it drops off of the host and spins an elongate whitish 



ROSS: EMBIA. BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 



33 



cocoon within a gallery, and pupates. Cocoons of 
some species are coated with debris, which must have 
been added by an embiid because they characteristi- 
cally eject or cover foreign objects encountered in 
their galleries. The cocoon is attached to the gallery 
wall near the shrivelled, dead body of the host. 

The sclerogibbids were revised by Richards ( 1 939) 
and there have been more recent papers dealing with 
the systematics and biology of several species. 
Scores of host-associated new species await descrip- 
tion in California Academy of Sciences as a result of 
my fieldwork and culturing activity. I have had no 
success in continuous rearing of the wasps in labora- 
tory cultures but this may be due to absence in a 
culture of both sexes at the same time to insure mat- 
ing. Shetlar (1973) succeeded in getting unmated fe- 
males of Pwbethylus schwarzi to reproduce, but, of 
course, the progeny were males. In another example, 
fifteen males, but no females, emerged from a cul- 
ture of Archembia n. sp., from Santa Catarina state, 
Brazil. 




Figure 39. Universal appearance of female sclerogibbids. 
Note peculiar enlarged forelegs. Superficially, species dif- 
fer only in size and coloration. Body length 3.0 mm. 



Sclerogibbids are themselves parasitized by 
wasps of the chalcidoid family Perilampidae, as evi- 
denced by the emergence of five Perilampus 
philembia Burks from as many cocoons of a sclerogib- 
bid parasitoid of Archembia n. sp. in Tingo Maria, 
Peru. More recently, apparently the same species 
emerged in a large culture of Gibocercus n. sp. from 
Ecuador's Rio Napo region. A pupa of the Tingo 
Maria sclerogibbid also yielded a male chalcidoid of 
the genus Mondontemerus (Torymidae). 

It is possible that sclerogibbids may be attacked 
by their hosts. Three large males collected in a cul- 
ture of Neorhagadochir Ross from Nicaragua suf- 
fered extensive loss of antennal segments and this 




Figure 40. Single sclerogibbid wasp larva feeding on adult 
female of a new genus and species of Anisembiidae occur- 
ring in upper Amazon of Brazil and Peru. 



I S 




M 



\ « 



.^.^i 



Figure 41. Mature sclerogibid larvae attached to adult 
female of Archembia n. sp. Tingo Maria. Peru. 



34 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



could only have resulted from nibbling by the host 
embiids. 

Another type of ectophagy involves larvae of 
small cecidomyid gnats of the genera Feltiella, or 
Lestodiplosis (Ross 1944:491 ). These are similar to 
sclerogibbid larvae both in appearance and method 
of feeding. They were encountered on one occasion 
only, in northern Florida, on Diradius vandykei 
(Ross). 

ENDOPARASITOIDS: Embiids, at least in 
Neotropical regions, apparently are hosts to braconid 
wasps of the genus Sericobracon ( Doryctinae). Stud- 
ies of such wasps were conducted by Scott R. Shaw 
and Janice S. Edgerly (1985). 

Embiidina are hosts to larvae of unusual tachinid 
flies as evidenced by collections I have made in widely 
separated geographic regions. Only two of the sev- 
eral new species have been described. Rossimyiopsis 
whiteheadi Mesnil, 1953, was reared from 
Apterembia cercocyrta (Krauss) in South Africa and 
E. I. Schlinger and I reared another series, Pet-umyia 
embiaphaga Arnaud (1963), from Clothoda 
longicauda Ross in Tingo Maria, Peru. Series I reared 
from various embiid hosts occurring in Central Amer- 
ica, Africa and tropical Asia await study. These ta- 
chinids are small, averaging about 3 mm in length; 
with shiny, not densely setose, black bodies, and usu- 



ally smoky wings. As they walk they rotate their 
wings. 

EGG PARASITOIDS: Tiny wasps developing 
within embiid eggs belong to genera Embidobia 
Ashmead and Palaeogiyon Masner, tribe Embidobi- 
ini, family Scelionidae. They occur almost through- 
out the range of Embiidina. The writer has reared 
and preserved numerous host-associated lots of spec- 
imens which should represent many new species. This 
collection is being studied by L. Masner. The fe- 
males of some species are pale ferrugineous and of- 
ten apterous or subalate (brachypterous). Successive 
generations can be reared in embiid cultures. At times 
only one sex of the wasps, usually males, appear in a 
particular culture. 

The tendency of most embiid species to coat their 
eggs with a hardened paste of chewed debris and their 
feces, may reduce oviposition by the wasps. Guard- 
ing by the parent female may also protect a large per- 
centage of the eggs in a mass (Fig. 42). 

PATHOLOGICAL HAZARDS: Disease epi- 
demics may weaken and even kill all individuals in a 
culture. It is assumed that such diseases also occur in 
nature but probably do not have such a catastrophic 
effect due to scattered occurrence of host colonies 
and consequent reduced contagion. 




Figure 42. Adult female of Cibocercus n. sp. from Ecuador resting on her mass of eggs 
covered with layers of silk. Note egg parasite approaching from her rear. 



ROSS: EMBIA. BIOSYSTEMATICS OFTHE ORDER EMBIIDINA, PART2 



35 



No studies specific to embiids have been made 
of these diseases but one can speculate that they are 
caused by viral, bacterial and fungal pathogens not 
necessarily restricted to embiids. In some cases ac- 
tivity of the victim will slow and eventually cease, 
the body turning reddish as its tissues liquefy. In 
cases of fungal infection, white mycelia begin to 
outline somites and sclerites of the embiid victim 
and later its entire body becomes a fuzzy, moldy mass. 
Gerald M. Thomas of the University of California 
identified the pathogen of one such epidemic, in a 
culture of an oligotomid from northwest Thailand, 
as Beauveria bassiana. He commented that this is 
the most commonly occurring insect pathogenic fun- 
gus in the world, and has a very wide host range on 
terrestrial and aquatic arthropods. 

The most important epidemics, however, are 
caused by sporozoan parasites of the gsntrd-Adelina 
(A. transita according to J. P. Kramer, pers. com.), 
Gregarina, and Diplocystis. An infected culture 



exhibits gradually reduced vitality, no new galleries 
are spun and eventually all the occupants die. Stefani 
(1959, 1960) made special studies of Diplocystis 
clerci parasitizing Haploembia solieri in the Medi- 
terranean region. He noted that the protozoa may 
damage sperm and weaken all males and thereby 
limit a species' reproduction to a residual minority 
of females capable of reproducing parthenogeneti- 
cally. Perhaps the consistently parthenogenetic form 
of H. solieri developed this way in populations on 
islands on the Tyrrhenean Sea on which all males 
had been exterminated by repeated epidemics (see 
section on parthenogenesis for details). 

Undoubtedly, numerous other microorganisms 
infect embiids throughout the order's range; how- 
ever, the only important investigations to date, are 
limited to Sardinian hosts studied by Stefani who 
also reported the parasitic nematode, Hexamennis, 
in two species of Embia. 





Figure 43. Culturing is a way of studying embiid biology and securing specimens for study. Those from Australia were 
maintained in my home laboratory. Each (left) had an associated jar (right) in which "harvested" series were preserved 
in alcohol jars. 








Figure 44 Left, a gallon-sized culture (cover removed) 
produced hundreds of adult specimens of Dihybocerciis 
n. sp. from Zambia. These half-grown, orange-and-black 
nymphs mimic poisonous paederine staphylinid beetles. 



36 



OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149 



Literature Cited 

Arnaud. P. H. 1963. Perumyia embiaphaga, a new 
genus and species of Neotropical Tachinidae 
(Diptera) parasitic on Embioptera. Amer. Mus. 
Novitates No. 2143:1-8, 14 figs. 

Carayon, J. 1974. Etudes sur les Hemipteres 
Plokiophilidae Ann. Soc. Ent. France (N.S.), 
10(3):449-525, 33 figs. 

Carrrera. M. 1947. Sohxto gentro Leptopteromyia 
Williston, 1908 (Diptera, Asilidae). Papeis 
Avulsos do Dept. Zool. Sec. Agric, Sao Paulo, 
8:89-96. 

Edgeriy, J. S. 1987a. Maternal behavior of a 
webspinner (Order Embiidina). Ecological 
Entomol. 12;1-11. 

Edgeriy, J. S. 1987b. Colony composition and some 
costs and benefits of facultively communal be- 
havior in Trinidadian webspinner, Clothoda 
uriclii (Embiidina: Clothodidae). Ann. Entomol. 
Soc. Amer. 80:29-34, 6 figs. 

Edgeriy, J. S. 1988. Maternal behavior of a 
webspinner (Order Embiidina): mother-nymph 
association. Ecological Entomol. 13:263-272. 

Edgeriy, J. S. 1994. Is group living an antipredator 
defense in a faculatively communal webspinner 
(Embiidina: Clothodidae)? Jour. Insect Behav- 
ior 7:135-147. 

Friederichs, K. 1934. Des Gemeinschaftsleben der 
Embiiden und Naheres zur Kenntnis der Arten. 
Archiv fur Naturg., N.F 3:405^M4, 18 figs. 

Mesnil. L. P. 1953. A new tachinid parasite of an 
Embiopteran. Proc. R. Entomol. Soc. London 
(B) 22:145-146, 1 fig., pi I. 

Myers, J. G. 1928. The first known embiophile, 
and a new Cuban embiid. Bull. Brooklyn 
Entomol. Soc. 23:87-90, 1 fig. 

Richards, O. W. 1939. The Bethylidae Subfamily 
Sclerogibbinae (Hymenoptra). Proc. R. Entomol. 
Soc. London (B) 8:211-223. 17 figs. 

Ross. E. S. 1940. A revision of the Embioptera of 
North America. .Ann. Entomol. Soc. Amer. 
33:629-676, 50 figs. 



Ross, E. S. 1956. A new genus of Embioptera from 
Baltic Amber. Mitt. Geol. Staatinst. Hamburg, 
25:76-81, 2 figs. 

Ross, E. S. 1957. The Embioptera of California. 
Bull. California Insect Survey 6:51-57, 7 figs. 

Ross, E. S. 1961. Parthenogenetic African 
Embioptera. Wasmann Jour. Biol. 18:297-304. 

Ross, E. S. 1966. A new species of Embioptera from 
the Galapagos Islands. Proc. California Acad. 
Sci. 34(12):499-504, 1 fig. 

Shaw, S. R. and J. S. Edgeriy. 1985. Anew 
brachonid genus (Hymenoptera) parasitizing 
webspinners (Embiidina) in Trinidad. Psyche 
92:505-511. 

Shetlar. D. J. 1973. A redescription and biology of 
Probethylus schwarzi Ashmead (Hymenoptera: 
Sclerogibbidae) with notes on related species. 
Entomol. News 84:205-210, 5 figs. 

Stefani, R. 1953a. Un particolare modo di 
accoppiamento negli Insetti Embiotteri. 
Rediconti Accad. Nazionale, Lincei (8) 14:544— 
549, 1 fig. 

Stefani, R 1953b. La fisiologia dell'accopiamento 
in Haploembia solieri Ram. (Embioptera: 
Oligotomidae). Rediconti Accad. Nazionale, 
Lincei (8) 15:211-216, 1 fig. 

Stefani, R. 1956. II problema della partenogenesi 
in Haploembia solieri Ramb. (Embioptera: 
Oligotomidae) Atti Accad. Nazionale, Lincei (8) 
5:127-201, 58 figs., pi. 1-2. 

Stefani, R. 1959. Dati sulla partenogenesi 
occidentali negli Embiotteri della fauna 
mediterrane. Rendiconte Accad. Nazionale 
Lincei, Classe Scienze fisiche matematiche 
naturali (8)25:622-625. 

Stefani, R. 1960. I rapporte tra parassitosi sterilita 
maschile partenogenesi accidentale in popolazio 
naturali di Haploembia solieri Ram. 
Anfigonica. RevistaParassitologia 2 1:277-287, 
4 figs. 

Stefani, R. 1961. La citologia della partenogenesi 
in due nuovi Embiotteri dell'Africa tropicale. 
Atti Accad. Nazionale Lincei (8) 30:254-256, 
Ipl. 



SMtTHSONIAN INSTTTUTION UBRARIEB 




3 9088 01302 6422